Skip to content

Instantly share code, notes, and snippets.

Show Gist options
  • Select an option

  • Save brett-petrusek/4d731beb1f440dc008f3abb1765eb0a6 to your computer and use it in GitHub Desktop.

Select an option

Save brett-petrusek/4d731beb1f440dc008f3abb1765eb0a6 to your computer and use it in GitHub Desktop.
roadmap and milestone planning

Professional Services Proposal - Discovery & Outcome Prioritization (SOP 1 of 4)

Overview

This is the first SOP in a 4-part series for creating fixed-price, milestone-based professional services project proposals. This SOP focuses on understanding the client's needs, extracting business outcomes from the Product Requirements Document (PRD), and prioritizing those outcomes based on strategic value and implementation feasibility.

SOP Sequence:

  1. Discovery & Outcome Prioritization ← You are here
  2. Milestone Scoping & Task Decomposition
  3. Estimation & Risk Analysis
  4. Pricing & Proposal Generation

This SOP analyzes the PRD and business context to identify the highest-priority business outcome that should drive the MVP/Phase 1 milestone definition in SOP 2.

Parameters

  • prd_document (required): The Product Requirements Document detailing the desired solution, features, and business outcomes.
  • business_context (optional): Additional business context including market conditions, competitive landscape, user research, strategic goals, organizational constraints, and client organization details.
  • timeline_constraints (optional): Hard deadlines, market windows, business events, or timing requirements that impact delivery.
  • budget_range (optional): Client's budget expectations or constraints.
  • success_metrics (optional): Existing KPIs, success criteria, or client's definition of success for project evaluation.
  • risk_tolerance (optional, default: "moderate"): Organization's appetite for risk - "conservative", "moderate", or "aggressive".
  • change_readiness (optional): Information about organizational change management capacity, user readiness for new solutions, and stakeholder engagement.
  • technology_stack (optional): Preferred or required technology choices including:
    • Frontend framework (e.g., React, Vue, Angular, Next.js)
    • Backend framework/language (e.g., Node.js/Express, Python/Django, .NET, Java/Spring)
    • Database (e.g., PostgreSQL, MySQL, MongoDB, DynamoDB)
    • Mobile (e.g., React Native, Flutter, Swift/Kotlin native)
    • Other key technologies (e.g., GraphQL vs REST, real-time requirements)
  • hosting_environment (optional): Hosting and infrastructure preferences including:
    • Cloud provider (e.g., AWS, Azure, GCP, DigitalOcean)
    • Deployment model (e.g., PaaS like Heroku/Vercel, containerized with ECS/Kubernetes, serverless, VPS)
    • CI/CD preferences (e.g., GitHub Actions, GitLab CI, Jenkins)
    • Infrastructure requirements (e.g., high availability, multi-region, compliance needs)
  • existing_systems (optional): Existing systems that must be integrated with, including:
    • System names and purposes (e.g., "Salesforce CRM", "SAP ERP", "Legacy inventory database")
    • Integration points and APIs available
    • Data sync requirements
    • Authentication/SSO requirements
  • technical_constraints (optional): Technology constraints including:
    • Must-use technologies (client/organizational standards)
    • Must-avoid technologies (licensing, security, organizational policies)
    • Team skill constraints (technologies team is proficient/not proficient in)
    • Compliance requirements (HIPAA, SOC2, GDPR, etc.)
  • working_directory (optional, default: "./proposal-outputs"): Directory path where all output artifacts will be saved.

Constraints for parameter acquisition:

  • You MUST ask for all required parameters upfront in a single prompt rather than one at a time
  • You MUST support multiple input methods for documents including:
    • Direct input: Text provided directly in the conversation
    • File path: Path to a local document file
    • URL: Link to a document
    • Other methods: You SHOULD be open to other ways the user might want to provide documents
  • You MUST use appropriate tools to access content based on the input method
  • You MUST prompt the user for technical context if not provided:
    • "What technology stack are you planning to use or do you have preferences? (Frontend, Backend, Database, Mobile)"
    • "What hosting environment are you targeting? (Cloud provider, deployment model)"
    • "Are there existing systems this needs to integrate with?"
    • "Are there any technical constraints? (Must-use/avoid technologies, compliance requirements)"
    • If user says "not sure" or "you recommend", document that technology choices will be made during Phase 1 design
  • You MUST confirm successful acquisition of all parameters before proceeding
  • You MUST create the working directory if it doesn't exist
  • You MUST save the input PRD document to "{working_directory}/prd-source.md" for reference in subsequent SOPs
  • You MUST save technical context to "{working_directory}/technical-context.md" for reference in subsequent SOPs (even if minimal/TBD)

Steps

1. Parse and Understand the PRD

Extract and categorize all elements from the PRD into a structured format.

Constraints:

  • You MUST identify and extract the following elements from the PRD:
    • Problem statement and user pain points
    • Proposed solution and key features
    • Target users and personas
    • Business goals and desired outcomes
    • Success metrics and KPIs
    • Technical requirements and constraints
    • Dependencies and assumptions
  • You MUST distinguish between explicitly stated business outcomes and implied outcomes
  • You MUST identify and flag any gaps or ambiguities in the PRD
  • You MUST flag any missing critical information that would affect prioritization because incomplete understanding leads to poor scoping decisions
  • You MUST create a structured breakdown with the following sections:
    • Problem & Pain Points
    • Proposed Solution Overview
    • Feature Catalog (complete list of all features mentioned)
    • Target Users & Personas
    • Business Goals & Outcomes
    • Success Metrics
    • Technical Requirements
    • Dependencies & Assumptions
    • Gaps & Ambiguities Identified
  • You MUST include explicit reasoning for interpretations and inferences:
    • For each implied outcome: Explain what PRD evidence led to this inference
    • For each gap identified: Explain why this information is critical
    • For each ambiguity: Explain the possible interpretations and which one was chosen (if any)
    • For feature categorization: Explain the logic used to organize features
  • You MUST save the structured breakdown to "{working_directory}/prd-breakdown-draft.md" (with "-draft" suffix)
  • You MUST include a Review Summary at the top of the artifact (see Review Summary Template below)
  • After human approval, You MUST rename the file from "prd-breakdown-draft.md" to "prd-breakdown.md"

Review Summary Requirements for prd-breakdown.md:

  • Key Findings: Feature count, primary problem, target users, stated vs. implied outcomes
  • Red Flags: Missing requirements, ambiguous features, conflicting information, gaps in success metrics
  • Critical Decisions: How implied outcomes were inferred, how ambiguities were interpreted
  • Critical Assumptions: Assumptions made about missing information
  • Approval Checklist: Verify feature catalog is complete, verify outcomes are correctly extracted, verify gaps are identified, verify no major misinterpretations

Human Review Gate: After generating prd-breakdown-draft.md, you MUST:

  • Present the Review Summary to the user
  • Ask the user to review the draft artifact
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft artifact and re-present for approval
  • Once approved, rename "prd-breakdown-draft.md" to "prd-breakdown.md"
  • Only proceed to Step 2 after the artifact is approved and renamed

2. Document Technical Context and Constraints

Capture technology stack preferences, hosting environment, existing systems, and technical constraints to inform solution design and estimation.

Constraints:

  • You MUST create a technical context document even if information is minimal or TBD
  • You MUST document the following (mark as "TBD - Will be determined in Phase 1 design" if not provided):
    • Technology Stack Preferences:
      • Frontend framework and rationale (if specified)
      • Backend framework/language and rationale (if specified)
      • Database choice and rationale (if specified)
      • Mobile technology (if applicable)
      • API architecture (REST, GraphQL, etc.)
    • Hosting & Infrastructure:
      • Cloud provider or hosting platform
      • Deployment model (PaaS, containerized, serverless, etc.)
      • CI/CD approach
      • Scalability and availability requirements
    • Existing Systems & Integrations:
      • List of systems to integrate with
      • Integration requirements (API types, data sync, auth)
      • Integration complexity assessment
    • Technical Constraints:
      • Must-use technologies (organizational standards, client requirements)
      • Must-avoid technologies (licensing, security, policies)
      • Compliance requirements (HIPAA, SOC2, GDPR, PCI-DSS, etc.)
      • Security requirements
  • You MUST assess how technical choices affect the project:
    • Team Familiarity: Is team experienced with chosen technologies? (affects risk and estimation)
    • Technology Maturity: Are technologies proven and stable? (affects risk)
    • Integration Complexity: How complex are the integrations? (affects effort and risk)
    • Learning Curve: Will team need to learn new technologies? (affects timeline and risk)
  • You MUST include explicit reasoning for technical assessments:
    • For team familiarity: Explain evidence of experience or lack thereof
    • For technology maturity: Cite evidence (widely adopted, cutting-edge, beta, etc.)
    • For integration complexity: Explain what makes integrations simple or complex
    • For learning curve: Estimate learning time needed (if applicable)
  • You MUST flag technical decisions that need to be made in Phase 1:
    • List decisions that are TBD (e.g., "Database choice between PostgreSQL and MongoDB")
    • Recommend including design/architecture work in Phase 1 if many decisions are TBD
    • Estimate time needed for technical design work (typically 1-2 weeks of Phase 1)
  • You MUST save technical context to "{working_directory}/technical-context-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename from "technical-context-draft.md" to "technical-context.md"

Review Summary Requirements for technical-context.md:

  • Key Findings: Technology decisions made vs. TBD, team familiarity level, integration complexity, compliance requirements
  • Red Flags: Many TBD decisions (increases Phase 1 design effort), unfamiliar technologies (increases risk), complex integrations, compliance requirements not addressed, must-use technologies team doesn't know
  • Critical Decisions: Technology choices that significantly affect design and estimation
  • Critical Assumptions: Technology assumptions that affect feasibility and effort
  • Approval Checklist: Verify technology choices are appropriate, verify team can execute on chosen stack, verify integrations are feasible, verify compliance requirements are identified, verify TBD items are flagged for Phase 1

Human Review Gate: After generating technical-context-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight what's decided vs. TBD
  • Flag any technology risks or unfamiliarity
  • Ask the user to validate technical context
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft and re-present
  • Once approved, rename to "technical-context.md"
  • Only proceed to Step 3 after approved and renamed

3. Extract and Categorize Business Outcomes

Identify all potential business outcomes and map them to measurable success criteria.

Constraints:

  • You MUST list all business outcomes mentioned or implied in the PRD
  • You MUST categorize outcomes by type:
    • Revenue outcomes: Increase sales, reduce churn, enable new revenue streams, improve pricing power
    • Efficiency outcomes: Reduce operational costs, improve productivity, automate processes, reduce manual effort
    • User experience outcomes: Increase satisfaction, improve engagement, reduce friction, enhance retention
    • Risk mitigation outcomes: Improve security, ensure compliance, reduce technical debt, business continuity
    • Market positioning outcomes: Competitive advantage, market expansion, brand strengthening, thought leadership
    • Strategic outcomes: Capability building, organizational transformation, data-driven decision making
  • You MUST define measurable success criteria for each outcome including:
    • Quantitative metrics where possible (e.g., "Reduce support tickets by 30%")
    • Qualitative indicators for outcomes that are harder to quantify
    • Baseline measurements if provided in the PRD
    • Target measurements that would indicate success
  • You MUST identify the primary stakeholders who care about each outcome (e.g., CEO, CFO, Sales VP, Customer Success, End Users)
  • You MUST estimate the potential impact magnitude for each outcome:
    • High Impact: Significant effect on revenue, costs, or strategic goals (>$500k annual value or >20% improvement)
    • Medium Impact: Noticeable improvement to operations or user experience ($100k-$500k annual value or 10-20% improvement)
    • Low Impact: Incremental benefit or nice-to-have (<$100k annual value or <10% improvement)
    • Note: Adjust these thresholds based on client size and context
  • You MUST assess measurability for each outcome:
    • Highly Measurable: Clear quantitative metrics, easy to track
    • Moderately Measurable: Mix of quantitative and qualitative indicators
    • Difficult to Measure: Primarily qualitative, subjective assessment required
  • You MUST include both quantitative and qualitative success metrics
  • You MUST identify any outcomes that lack clear success criteria and flag them for clarification
  • You MUST include explicit reasoning for categorization and assessment decisions:
    • For each outcome categorization: Explain why it was placed in that category (e.g., "Categorized as Revenue outcome because it directly enables new sales through...")
    • For each impact magnitude assessment: Explain the calculation or logic (e.g., "Estimated High Impact ($600k annual) based on 30% ticket reduction × 2 FTE @ $150k/year...")
    • For each stakeholder identification: Explain why they care about this outcome
    • For each measurability assessment: Explain what makes it easy/hard to measure
  • You MUST save the business outcomes analysis to "{working_directory}/business-outcomes-draft.md" (with "-draft" suffix)
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "business-outcomes-draft.md" to "business-outcomes.md"

Review Summary Requirements for business-outcomes.md:

  • Key Findings: Total outcome count by category, primary outcome types, stakeholder distribution
  • Red Flags: Outcomes without measurable criteria, conflicting outcomes, unrealistic success targets, outcomes that can't be validated
  • Critical Decisions: How outcomes were categorized, how impact magnitude was assessed
  • Critical Assumptions: Baseline metrics assumptions, impact value calculations
  • Approval Checklist: Verify all outcomes are captured, verify success criteria are realistic and measurable, verify stakeholder mapping is correct, verify no outcomes are impossible to validate

Human Review Gate: After generating business-outcomes-draft.md, you MUST:

  • Present the Review Summary to the user
  • Ask the user to review the draft artifact
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft artifact and re-present for approval
  • Once approved, rename "business-outcomes-draft.md" to "business-outcomes.md"
  • Only proceed to Step 4 after the artifact is approved and renamed

4. Analyze Business Context and Strategic Fit

Evaluate the broader business context to understand which outcomes align with strategic priorities.

Constraints:

  • You MUST analyze the following context factors for each business outcome:
    • Strategic alignment: How each outcome supports organizational strategy, mission, and long-term goals
    • Market timing: Windows of opportunity, competitive pressures, first-mover advantages, market trends
    • User urgency: Critical pain points vs. nice-to-haves, severity of current problems, user demand intensity
    • Resource availability: Team capacity, capability constraints, budget limitations, technology constraints
    • Technical feasibility: Technology maturity, integration complexity, known technical challenges
    • Risk factors: Technical risks, market risks, organizational risks, dependency risks
  • You MUST identify any external factors that create urgency including:
    • Regulatory deadlines or compliance requirements
    • Competitive threats or market disruption
    • Market windows or seasonal timing
    • Funding milestones or investor expectations
    • Contract renewals or customer commitments
  • You MUST assess organizational readiness for change in each outcome area:
    • Stakeholder buy-in and support
    • Change management capacity
    • Process maturity and readiness
    • Cultural alignment
    • Resource availability for adoption
  • You MUST identify potential blockers or dependencies for each outcome:
    • Technical dependencies (infrastructure, integrations, third-party services)
    • Organizational dependencies (other initiatives, decision makers, resource allocation)
    • External dependencies (vendors, partners, regulatory approval)
    • Sequential dependencies (prerequisites that must be completed first)
  • You MUST assess the client's risk tolerance and how it affects outcome feasibility:
    • Conservative clients: Favor proven approaches, incremental change, low technical risk
    • Moderate clients: Balance innovation with pragmatism, accept managed risk
    • Aggressive clients: Willing to take bigger bets, embrace novel approaches, higher risk tolerance
  • You SHOULD identify potential synergies between outcomes (where pursuing one outcome enables or accelerates another)
  • You MUST include explicit reasoning for all assessments:
    • For each strategic alignment rating: Explain how the outcome supports specific strategic goals
    • For each urgency assessment: Explain what creates the urgency (market pressure, deadline, pain severity, etc.)
    • For each feasibility concern: Explain the specific constraints or risks that affect feasibility
    • For each blocker identified: Explain why it's a blocker and what would be needed to remove it
    • For risk tolerance interpretation: Explain what evidence led to the assessment
  • You MUST document context analysis in "{working_directory}/business-context-analysis-draft.md" (with "-draft" suffix)
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "business-context-analysis-draft.md" to "business-context-analysis.md"
  • You MUST NOT ignore organizational or cultural constraints because these often determine actual feasibility more than technical factors

Review Summary Requirements for business-context-analysis.md:

  • Key Findings: Primary strategic drivers, most urgent outcomes, major constraints identified, risk tolerance assessment
  • Red Flags: Outcomes with major blockers, unrealistic timelines given constraints, resource availability concerns, organizational readiness gaps, cultural resistance risks
  • Critical Decisions: How strategic alignment was assessed, risk tolerance interpretation
  • Critical Assumptions: Market timing assumptions, stakeholder support assumptions, resource availability assumptions
  • Approval Checklist: Verify strategic priorities are correctly understood, verify constraints are accurately captured, verify urgency factors are valid, verify organizational readiness assessment is realistic

Human Review Gate: After generating business-context-analysis-draft.md, you MUST:

  • Present the Review Summary to the user
  • Ask the user to review the draft artifact
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft artifact and re-present for approval
  • Once approved, rename "business-context-analysis-draft.md" to "business-context-analysis.md"
  • Only proceed to Step 5 after the artifact is approved and renamed

5. Prioritize Business Outcomes

Rank business outcomes using a multi-dimensional prioritization framework to identify the top outcome that should drive MVP/Phase 1.

Constraints:

  • You MUST evaluate each business outcome across these six dimensions (1-5 scale, 5 = best):
    • Strategic Value (1-5): Alignment with business strategy and long-term goals
      • 5: Mission-critical, core to business strategy
      • 4: Strongly aligned with strategic priorities
      • 3: Supports strategy, not central
      • 2: Tangential to strategy
      • 1: Minimal strategic connection
    • User Impact (1-5): Magnitude of improvement to user experience or pain point resolution
      • 5: Solves critical pain point affecting all users
      • 4: Major improvement to frequent user activities
      • 3: Moderate improvement to some user workflows
      • 2: Minor convenience for subset of users
      • 1: Negligible user-facing impact
    • Revenue Potential (1-5): Direct or indirect revenue impact
      • 5: >$1M annual revenue impact or enables major deals
      • 4: $500k-$1M annual revenue impact
      • 3: $100k-$500k annual revenue impact
      • 2: $50k-$100k annual revenue impact
      • 1: <$50k annual revenue impact
      • Note: Adjust thresholds based on client size
    • Time to Value (1-5): How quickly benefits can be realized (5 = fastest)
      • 5: Value realized within 4-8 weeks
      • 4: Value realized within 8-12 weeks
      • 3: Value realized within 12-16 weeks
      • 2: Value realized within 16-24 weeks
      • 1: Value realized after 6+ months
    • Implementation Risk (1-5): Technical and organizational risk (5 = lowest risk)
      • 5: Proven approach, minimal dependencies, low complexity
      • 4: Mostly familiar territory, manageable risks
      • 3: Some novel aspects, moderate complexity
      • 2: Significant unknowns, high complexity
      • 1: Highly uncertain, experimental, many dependencies
    • Resource Efficiency (1-5): Return on investment of development effort (5 = best ROI)
      • 5: Massive value for minimal effort
      • 4: Strong ROI, efficient value delivery
      • 3: Reasonable ROI, effort matches value
      • 2: Marginal ROI, high effort for moderate value
      • 1: Poor ROI, disproportionate effort
  • You MUST calculate a weighted priority score for each outcome based on the organization's risk tolerance:
    • Conservative weighting:
      • Time to Value: 25%
      • Implementation Risk: 25%
      • Strategic Value: 20%
      • Revenue Potential: 15%
      • User Impact: 10%
      • Resource Efficiency: 5%
    • Moderate weighting (balanced):
      • Strategic Value: 20%
      • Revenue Potential: 20%
      • User Impact: 20%
      • Time to Value: 15%
      • Implementation Risk: 15%
      • Resource Efficiency: 10%
    • Aggressive weighting:
      • Strategic Value: 30%
      • Revenue Potential: 30%
      • User Impact: 15%
      • Resource Efficiency: 10%
      • Time to Value: 10%
      • Implementation Risk: 5%
  • You MUST calculate the weighted score for each outcome: Sum(Dimension Score × Weight)
  • You MUST rank all outcomes by their weighted priority score (highest to lowest)
  • You MUST identify the top 3 highest-priority outcomes
  • You MUST clearly document the rationale for the top-ranked outcome including:
    • Why it scored highest across the dimensions
    • What makes it the best choice for Phase 1/MVP focus
    • How it compares to the other top contenders
    • Any close calls or tie-breaking decisions made
  • You MUST flag any outcomes that scored very close to each other (within 0.3 points) as "close calls" requiring human judgment
  • You MUST create a visual representation of the prioritization using a markdown table:
    | Outcome | Strategic Value | User Impact | Revenue Potential | Time to Value | Implementation Risk | Resource Efficiency | Weighted Score | Rank |
    |---------|----------------|-------------|-------------------|---------------|---------------------|---------------------|----------------|------|
    | ...     | X/5            | X/5         | X/5               | X/5           | X/5                 | X/5                 | X.XX           | 1    |
    
  • You SHOULD create a value/effort matrix visualization in markdown table format showing outcomes plotted by value (Strategic Value + Revenue Potential + User Impact) vs. effort (inverse of Resource Efficiency and Implementation Risk)
  • You MUST include detailed reasoning for ALL scoring decisions:
    • For EACH outcome's EACH dimension score: Provide 1-2 sentence explanation of why that specific score was assigned
    • For example: "Strategic Value: 5/5 - This outcome directly addresses the company's #1 strategic priority of reducing customer churn, mentioned explicitly in the business context as 'mission-critical'"
    • For revenue potential: Show the calculation or estimation logic (e.g., "$400k annual = 50% stock-out reduction × $800k current annual stock-out costs")
    • For time to value: Explain the timeline estimate and key assumptions
    • For implementation risk: Identify the specific risks that drove the score
    • For resource efficiency: Compare effort to value delivered
  • You MUST explain the weighting choice:
    • Why conservative/moderate/aggressive was selected based on client profile
    • How the weighting reflects their priorities and constraints
  • You MUST explain any tie-breaking decisions:
    • What factors were considered to break ties between close-scoring outcomes
    • Why one outcome was prioritized over another when scores were similar
  • You MUST save the prioritization analysis to "{working_directory}/outcome-prioritization-draft.md" (with "-draft" suffix)
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "outcome-prioritization-draft.md" to "outcome-prioritization.md"

Review Summary Requirements for outcome-prioritization.md:

  • Key Findings: Top 3 outcomes with scores, winning outcome and why, close calls that need judgment, weighting approach used
  • Red Flags: Controversial rankings, outcomes that scored well on some dimensions but poorly on others, very close scores requiring tiebreaker, outcomes where scoring seems counterintuitive
  • Critical Decisions: Weighting rationale, how ties were broken, dimension score justifications for top outcomes
  • Critical Assumptions: Revenue impact calculations, risk assessments, time-to-value estimates
  • Approval Checklist: Verify top outcome makes intuitive sense, verify scoring reflects business priorities, verify no obviously better outcome was ranked lower, verify weighting is appropriate for client's risk tolerance, verify close calls are identified

Human Review Gate: After generating outcome-prioritization-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the top-ranked outcome and the rationale
  • Call out any close calls or controversial rankings
  • Ask the user to review and confirm agreement with the prioritization
  • Wait for explicit approval before proceeding
  • If the user disagrees with the ranking, discuss the reasoning and adjust scoring/weighting as needed, then update the draft and re-present for approval
  • Once approved, rename "outcome-prioritization-draft.md" to "outcome-prioritization.md"
  • Only proceed to Step 6 (generate SOP 2 execution prompt) after the artifact is approved and renamed

6. Generate SOP 2 Execution Prompt

Create the complete execution prompt for the next SOP in the sequence, including all necessary parameters and file paths.

Constraints:

  • You MUST generate a prompt that includes:
    • Clear instruction to execute SOP 2 (Milestone Scoping & Task Decomposition)
    • All file paths to artifacts generated in this SOP
    • The original prd_document path
    • Any optional parameters that should carry forward (timeline_constraints, budget_range, etc.)
    • The working_directory path
  • You MUST format the prompt as a ready-to-use execution command
  • You MUST include a brief summary of what was accomplished in SOP 1 to provide context
  • You MUST list the top-priority outcome that will drive Phase 1/MVP definition
  • You SHOULD include any important caveats, assumptions, or notes that the next SOP should be aware of

Prompt Template:

✅ SOP 1 (Discovery & Outcome Prioritization) completed successfully.

## Summary of SOP 1 Outputs:
- **Top Priority Outcome**: [Outcome name and brief description]
- **Priority Score**: [X.XX/5.00]
- **Key Rationale**: [1-2 sentence summary of why this outcome was prioritized]
- **Technical Context**: [Decided vs. TBD summary - e.g., "Stack: React/Node/Postgres on AWS (decided), Mobile: TBD in Phase 1"]
- **Artifacts Generated**:
  - prd-breakdown.md
  - technical-context.md
  - business-outcomes.md
  - business-context-analysis.md
  - outcome-prioritization.md

## Next Step:
Execute **SOP 2: Milestone Scoping & Task Decomposition** to define the MVP/Phase 1 scope and structure the remaining milestone phases.

---

## SOP 2 Execution Prompt:

Execute the Milestone Scoping & Task Decomposition SOP (proposalgen-sop-2-milestone-scoping.sop.md) with the following parameters:

**Required Parameters:**
- prd_document: "[path/to/prd-source.md]"
- prd_breakdown: "[path/to/prd-breakdown.md]"
- technical_context: "[path/to/technical-context.md]"
- business_outcomes: "[path/to/business-outcomes.md]"
- business_context_analysis: "[path/to/business-context-analysis.md]"
- outcome_prioritization: "[path/to/outcome-prioritization.md]"
- top_priority_outcome: "[Name of the #1 ranked outcome]"

**Optional Parameters:**
[Include any optional parameters that were provided in SOP 1]
- timeline_constraints: "[value if provided]"
- budget_range: "[value if provided]"
- resource_constraints: "[value if provided]"
- risk_tolerance: "[value, default: moderate]"
- working_directory: "[path/to/proposal-outputs]"

**Important Context for SOP 2:**
[Include any important notes, caveats, or assumptions that SOP 2 should be aware of]
- [Note 1]
- [Note 2]

Output:

  • You MUST present this prompt to the user
  • You MUST clearly indicate that SOP 1 is complete and they should execute the prompt above when ready to proceed
  • You SHOULD remind the user that they can modify any of the parameters before executing SOP 2 if needed

Review Summary Template

Every artifact generated in this SOP MUST include a Review Summary section at the very top, using this exact template:

# [Artifact Name]

## 📋 REVIEW SUMMARY
**Status**: Ready for Review
**Generated**: [ISO 8601 timestamp]
**Confidence Level**: [X/100] - [High (80-100) / Medium (60-79) / Low (40-59) / Very Low (<40)]

### 🎯 Key Findings
- [3-5 bullet points with the most important conclusions]
- [Key numbers, decisions, or outcomes]
- [What changed from inputs to outputs]

### ⚠️ RED FLAGS & SCRUTINY AREAS

**High Priority Issues:**
- [Things that are risky, uncertain, or could derail the project]
- [Assumptions that might be wrong]
- [Gaps in information or analysis]

**Moderate Concerns:**
- [Areas that need validation but aren't show-stoppers]
- [Ambiguities that should be clarified]

### 🔑 Critical Decisions & Assumptions

**Decisions Made:**
- [Key choices that affect downstream work]
- [Trade-offs and rationale]

**Critical Assumptions:**
- [Assumptions this analysis depends on]
- [What happens if these assumptions are wrong]

### 📊 Quantitative Summary
- [Key metrics, counts, totals relevant to this artifact]
- [Percentages, scores, confidence intervals]

### ✅ Approval Checklist
Before approving, verify:
- [ ] [Specific thing #1 to check - tailored to this artifact]
- [ ] [Specific thing #2 to check - tailored to this artifact]
- [ ] [Specific thing #3 to check - tailored to this artifact]
- [ ] No major elements overlooked
- [ ] Assumptions are reasonable and documented

**Action**: If approved, proceed to next step. If issues found, document feedback and request revisions.

---

## DETAILED ANALYSIS
[Full artifact content follows...]

Examples

Example Input

prd_document: "path/to/inventory-management-system-prd.md"

business_context: "Mid-sized manufacturing company, 200 employees. Currently using spreadsheets for inventory. Growth causing stock-outs and excess inventory. Manual processes consuming 2 FTEs. Risk-averse organization with limited IT staff. Need to prove value before full commitment."

timeline_constraints: "CEO wants to see progress before Q4 budget cycle (6 months out). Year-end inventory requires system operational by December."

budget_range: "$180k-$250k, prefer phased investment to spread cost"

success_metrics: "Reduce stock-outs by 50%, reduce excess inventory by 30%, reduce manual effort by 1.5 FTE, achieve inventory accuracy >95%"

risk_tolerance: "conservative"

change_readiness: "Operations team frustrated with current spreadsheet chaos and eager for improvement. Finance team concerned about cost. IT team has limited bandwidth for support."

working_directory: "./inventory-mgmt-proposal"

Example Output Summary (outcome-prioritization.md excerpt)

# Business Outcome Prioritization Analysis

## 📋 REVIEW SUMMARY
**Status**: Ready for Review
**Generated**: 2025-01-15T14:32:00Z
**Confidence Level**: 78/100 - Medium-High

### 🎯 Key Findings
- 7 distinct business outcomes identified from PRD and business context
- **Top Priority Outcome**: "Achieve accurate real-time inventory visibility to reduce stock-outs" (Score: 4.35/5.00)
- Conservative weighting applied based on risk-averse organizational profile
- Close call between top outcome (4.35) and #2 "Reduce manual effort and labor costs" (4.10) - only 0.25 point difference
- All outcomes have measurable success criteria; inventory accuracy is most quantifiable

### ⚠️ RED FLAGS & SCRUTINY AREAS

**High Priority Issues:**
- **Close scoring**: Top 2 outcomes scored within 0.25 points - prioritization could go either way depending on stakeholder input
- **Timeline tension**: Top outcome (inventory visibility) estimated at 10-12 weeks vs. labor reduction (8-10 weeks), but labor reduction scored lower on strategic value
- **Resource constraints**: IT bandwidth concerns may affect implementation risk scoring for integration-heavy outcomes

**Moderate Concerns:**
- Revenue potential scoring is somewhat speculative - based on rough $400k annual stock-out cost estimate from PRD
- Implementation risk for ERP integration (part of top outcome) may be underestimated if ERP API is poorly documented
- User impact scoring assumes operations team's enthusiasm translates to adoption - change management could be harder than expected

### 🔑 Critical Decisions & Assumptions

**Decisions Made:**
- Applied Conservative weighting (Time to Value: 25%, Implementation Risk: 25%) based on risk-averse culture and "need to prove value" requirement
- Prioritized inventory visibility over labor reduction despite labor reduction being faster to value, because inventory visibility has stronger strategic alignment and revenue impact
- Scored "automated reordering" lower due to high technical complexity and dependency on having accurate inventory data first

**Critical Assumptions:**
- Stock-out reduction directly correlates with inventory accuracy (assumed 95% accuracy → 50% stock-out reduction)
- $400k annual cost of stock-outs (from PRD) is accurate
- Operations team's eagerness for change means high user adoption likelihood
- ERP integration is feasible within moderate risk profile (scored 3/5 on Implementation Risk)
- 10-12 week timeline for inventory visibility is achievable given team constraints

### 📊 Quantitative Summary
- **Total Outcomes Evaluated**: 7
- **Outcome Categories**: 3 Revenue, 2 Efficiency, 1 User Experience, 1 Risk Mitigation
- **Top Score**: 4.35/5.00 (Inventory Visibility & Stock-out Reduction)
- **Lowest Score**: 2.45/5.00 (Advanced Analytics & Forecasting)
- **Close Calls (within 0.3 points)**: Outcomes #1 and #2 (0.25 point difference)
- **Weighting Applied**: Conservative (Risk-averse organization)

### ✅ Approval Checklist
Before approving, verify:
- [ ] Top outcome (Inventory Visibility) aligns with CEO's expectation to "see progress before Q4" and makes business sense
- [ ] Conservative weighting is appropriate given risk-averse culture (vs. moderate or aggressive)
- [ ] Close call between #1 and #2 has been reviewed - confirm #1 is the right choice or if #2 (labor reduction) should be prioritized instead
- [ ] Revenue potential estimates ($400k stock-out costs) are realistic
- [ ] Implementation risk scores (especially for ERP integration) reflect reality
- [ ] Timeline estimates (10-12 weeks for top outcome) are feasible
- [ ] No major elements overlooked
- [ ] Assumptions are reasonable and documented

**Action**: If approved, proceed to SOP 2. If #2 outcome (labor reduction) should be prioritized instead, provide feedback to adjust ranking.

---

## DETAILED ANALYSIS

### Prioritization Framework

**Risk Tolerance**: Conservative
**Weighting Applied**:
- Time to Value: 25%
- Implementation Risk: 25%
- Strategic Value: 20%
- Revenue Potential: 15%
- User Impact: 10%
- Resource Efficiency: 5%

### Outcome Rankings

| Rank | Outcome | Strategic Value | User Impact | Revenue Potential | Time to Value | Implementation Risk | Resource Efficiency | Weighted Score |
|------|---------|----------------|-------------|-------------------|---------------|---------------------|---------------------|----------------|
| 1 | Inventory Visibility & Stock-out Reduction | 5/5 | 4/5 | 5/5 | 3/5 | 3/5 | 4/5 | **4.35** |
| 2 | Reduce Manual Effort & Labor Costs | 4/5 | 5/5 | 3/5 | 4/5 | 4/5 | 4/5 | **4.10** |
| 3 | Reduce Excess Inventory & Carrying Costs | 4/5 | 3/5 | 4/5 | 3/5 | 3/5 | 3/5 | **3.50** |

### Detailed Outcome Analysis with Reasoning

#### Outcome #1: Inventory Visibility & Stock-out Reduction (Score: 4.35)

**Strategic Value: 5/5**
*Reasoning*: This outcome directly addresses the company's stated top strategic priority of improving inventory accuracy and reducing lost sales from stock-outs. The business context explicitly mentions "growth causing stock-outs" and CEO's requirement to "see progress before Q4 budget cycle." This is mission-critical to the business.

**User Impact: 4/5**
*Reasoning*: Affects all operations team members (primary users) who currently struggle with spreadsheet-based tracking. Addresses the severe pain point mentioned in context: "frustrated with current spreadsheet chaos." Not 5/5 because it primarily benefits internal operations team, not end customers directly.

**Revenue Potential: 5/5**
*Reasoning*: Success metrics state "Reduce stock-outs by 50%." Estimated $400k annual revenue impact based on: Current stock-out costs not explicitly stated in PRD, but 50% reduction in stock-outs for a 200-employee manufacturing company experiencing "growth causing stock-outs" conservatively estimated at $800k annual impact = $400k savings. Enables revenue that's currently being lost.

**Time to Value: 3/5**
*Reasoning*: Estimated at 10-12 weeks to achieve inventory visibility for initial SKU set. This is moderate speed - not fastest (would be 4-8 weeks for 5/5) but reasonable given the need for integration and user adoption. Timeline constrained by need to integrate with existing systems and train operations team.

**Implementation Risk: 3/5**
*Reasoning*: Moderate risk due to: (1) Need to integrate with existing ERP system (mentioned as dependency), (2) Data migration from spreadsheets to structured database, (3) User adoption required across operations team. However, inventory tracking is a well-understood problem domain with proven solutions, preventing this from being high-risk (1-2/5). IT team has "limited bandwidth" which adds to risk.

**Resource Efficiency: 4/5**
*Reasoning*: Strong ROI - estimated $400k annual benefit vs. estimated 10-12 weeks effort (roughly $80-100k investment based on typical 3-person team). Ratio of 4:1 or better annual return. Not 5/5 because ERP integration adds effort without direct user-facing value.

**Overall**: Top-ranked because it combines mission-critical strategic value (5/5) with strong revenue impact (5/5) while maintaining acceptable risk (3/5) - critical for conservative risk tolerance weighting.

---

#### Outcome #2: Reduce Manual Effort & Labor Costs (Score: 4.10)

**Strategic Value: 4/5**
*Reasoning*: Business context states manual processes are "consuming 2 FTEs" and success criteria includes "reduce manual effort by 1.5 FTE." This is a clear strategic priority but secondary to the stock-out problem based on context emphasis. Scored 4/5 instead of 5/5 because it's an efficiency play rather than the revenue-critical issue.

**User Impact: 5/5**
*Reasoning*: Directly reduces daily frustration for operations team doing manual data entry. Success metric of "1.5 FTE reduction" means massive impact on user daily experience. Operations team is "eager for improvement" per change readiness assessment. Maximum score because it eliminates repetitive manual work.

**Revenue Potential: 3/5**
*Reasoning*: $225k annual savings calculation: 1.5 FTE × $150k fully-loaded cost = $225k. This is moderate revenue impact - significant but less than the $400k from stock-out reduction. Indirect revenue impact through operational efficiency vs. direct revenue impact.

**Time to Value: 4/5**
*Reasoning*: Estimated 8-10 weeks - faster than inventory visibility outcome because labor reduction can be achieved through basic automation without complex ERP integration. Can show manual effort reduction quickly once basic data entry automation is in place.

**Implementation Risk: 4/5**
*Reasoning*: Lower risk than inventory visibility because: (1) Doesn't require ERP integration initially, (2) Proven automation patterns, (3) High user eagerness reduces adoption risk. Scored 4/5 instead of 5/5 because change management still needed for process changes.

**Resource Efficiency: 4/5**
*Reasoning*: Good ROI - $225k annual benefit vs. 8-10 weeks effort (~$60-80k investment) = 3:1 ratio. Slightly lower than Outcome #1 in absolute dollars but faster time to achieve.

**Overall**: Close second place (only 0.25 points behind #1). Loses to Outcome #1 primarily because Conservative weighting emphasizes Strategic Value (20%) and Implementation Risk (25%), where Outcome #1 has advantages. Under Aggressive weighting, these two would be much closer or potentially reversed.

[... analysis continues for other outcomes ...]

Troubleshooting

PRD Lacks Clear Business Outcomes

If the PRD focuses primarily on features without explicit business outcomes:

  • Infer business outcomes from the problem statement and user pain points
  • Ask clarifying questions about the business goals and success metrics using interview-style prompts
  • Document assumptions about intended outcomes and clearly mark them as "inferred" vs. "stated"
  • Flag these inferred outcomes in the Red Flags section of the Review Summary
  • Proceed with best-judgment outcome identification clearly marked as assumptions requiring validation
  • Recommend that the client provide additional context or validate the inferred outcomes

Business Context Is Limited or Missing

If insufficient business context is provided:

  • Request specific information about strategic priorities, market conditions, and constraints before proceeding
  • If user cannot provide context, make reasonable assumptions based on:
    • Industry norms and standard business drivers
    • Company size and maturity level
    • Problem severity indicated in the PRD
  • Document all assumptions prominently in the Review Summary Red Flags section
  • Use conservative estimates for impact and risk (when in doubt, score lower)
  • Build in flexibility for scope adjustment as more context emerges
  • Apply "Medium" confidence level and note that additional context could change prioritization

Multiple Outcomes Have Equal or Very Close Priority Scores

If several outcomes score within 0.3 points of each other:

  • Flag this as a "close call" in the Review Summary
  • Present all closely-scored outcomes to the user with detailed comparison
  • Look for tiebreaker factors:
    • Outcomes that enable or accelerate other outcomes (choose the enabler)
    • External forcing functions (deadlines, market windows, competitive threats)
    • Which outcome provides the strongest validation of core assumptions
    • Which outcome has lowest implementation risk if other factors are equal
  • Document the close decision and note that runner-up outcomes should be addressed in Phase 2
  • Recommend user input to make the final prioritization decision
  • Do NOT arbitrarily pick one - surface the decision to the human reviewer

Success Metrics Are Vague or Missing

If the PRD lacks specific, measurable success criteria:

  • Create reasonable proxy metrics based on the problem statement
  • Use industry benchmarks or standard metrics for similar outcomes
  • Mark these as "proposed" or "assumed" metrics requiring validation
  • Include specific questions in the Review Summary for the user to clarify
  • Flag in Red Flags section that success validation may be difficult
  • Recommend that the client define clear success metrics before proceeding to implementation
  • Consider whether lack of metrics should lower the priority score for that outcome

Client Risk Tolerance Is Unclear

If risk tolerance is not specified or ambiguous:

  • Default to "moderate" risk tolerance weighting
  • Present outcomes using all three weighting approaches (conservative, moderate, aggressive) to show how rankings change
  • Ask specific questions to clarify:
    • "Is the organization willing to take bigger bets for potentially higher returns?"
    • "Is proving value quickly more important than maximizing strategic impact?"
    • "How much appetite is there for novel/unproven approaches?"
  • Document the uncertainty in the Review Summary
  • Note that the prioritization may change based on confirmed risk tolerance

Timeline or Budget Constraints Conflict with Prioritization

If the top-priority outcome appears infeasible given stated constraints:

  • DO NOT artificially adjust scores to make outcomes fit constraints
  • Clearly document the conflict in the Review Summary Red Flags
  • Present the prioritization as-is based on business value
  • Flag that SOP 2 will need to address scope trade-offs or SOP 3 may reveal timeline/budget issues
  • Provide options for the user:
    • Prioritize a different outcome that fits constraints
    • Adjust constraints (extend timeline, increase budget)
    • Plan for a more limited Phase 1 scope that fits constraints
  • Let the user decide how to resolve the conflict before proceeding

User Disagrees with Outcome Prioritization

If the user reviews the prioritization and disagrees:

  • Ask clarifying questions to understand their reasoning
  • Identify which dimension scores they disagree with and why
  • Adjust scores based on their input (they have better business context)
  • Consider whether different weighting would better reflect priorities
  • Regenerate the prioritization with adjusted scores/weights
  • Document the changes made and the rationale in the artifact
  • Re-present for approval
  • Remember: The user is the expert on their business priorities - defer to their judgment

Professional Services Proposal - Milestone Scoping & Task Decomposition (SOP 2 of 4)

Overview

This is the second SOP in a 4-part series for creating fixed-price, milestone-based professional services project proposals. This SOP focuses on transforming the prioritized business outcomes into a structured milestone delivery plan with detailed task breakdowns.

SOP Sequence:

  1. Discovery & Outcome Prioritization
  2. Milestone Scoping & Task Decomposition ← You are here
  3. Estimation & Risk Analysis
  4. Pricing & Proposal Generation

This SOP takes the top-priority outcome from SOP 1 and defines a focused MVP/Phase 1 scope, structures the remaining work into logical milestone phases, and decomposes all work into estimable tasks. The output provides the foundation for effort estimation in SOP 3.

Critical Constraint: Phase 1/MVP MUST be scoped to be completable within 8-12 weeks (2-3 months) maximum. This is a hard constraint that drives all scoping decisions. If the minimum viable feature set to achieve the top-priority outcome exceeds 12 weeks, you must either descope further, narrow the user base, reduce feature depth, or recommend alternative approaches.

Parameters

  • prd_document (required): Path to the original PRD source file (from SOP 1).
  • prd_breakdown (required): Path to the structured PRD breakdown from SOP 1.
  • technical_context (required): Path to the technical context document from SOP 1 (includes technology stack, hosting, integrations, constraints).
  • business_outcomes (required): Path to the business outcomes analysis from SOP 1.
  • business_context_analysis (required): Path to the business context analysis from SOP 1.
  • outcome_prioritization (required): Path to the outcome prioritization analysis from SOP 1.
  • top_priority_outcome (required): Name/description of the #1 ranked business outcome that will drive Phase 1/MVP.
  • timeline_constraints (optional): Hard deadlines, market windows, or timing requirements.
  • budget_range (optional): Client's budget expectations or constraints.
  • resource_constraints (optional): Team size, skills, availability, and technology constraints.
  • risk_tolerance (optional, default: "moderate"): Organization's appetite for risk - "conservative", "moderate", or "aggressive".
  • working_directory (required): Directory path where all artifacts are saved (same as SOP 1).

Constraints for parameter acquisition:

  • You MUST ask for all required parameters upfront in a single prompt rather than one at a time
  • You MUST verify that all input files from SOP 1 exist and are readable
  • You MUST confirm successful acquisition of all parameters before proceeding
  • You SHOULD load and reference the SOP 1 artifacts to inform scoping decisions

Steps

1. Define MVP / Phase 1 Milestone

Identify the minimum valuable deliverable that achieves the highest-priority business outcome from SOP 1.

Constraints:

  • You MUST focus on the single highest-priority business outcome from SOP 1 (the top_priority_outcome parameter)
  • You MUST reference the technical context document to inform Phase 1 scoping:
    • If technology stack is decided: Use it to inform feature feasibility and task decomposition
    • If technology stack is TBD: Include solution design/architecture work in Phase 1 (typically 1-2 weeks)
    • If complex integrations exist: Consider integration complexity in scope and timeline
    • If team lacks familiarity with required technologies: Add learning curve buffer or recommend training
  • You MUST select features and capabilities that:
    • Directly enable the target business outcome
    • Can be delivered in 8-12 weeks maximum (2-3 months hard constraint for MVP/Phase 1)
    • Are technically feasible given the technology context (or include design work to determine feasibility)
    • Provide standalone value that can be deployed and used independently
    • Have manageable technical risk for fixed-price commitment
    • Demonstrate progress and build client confidence
  • You MUST ruthlessly descope Phase 1 if initial feature set exceeds 12 weeks - the 2-3 month timeline is a hard constraint, not a guideline
  • You MUST categorize all PRD features as:
    • Phase 1 Core: Essential for achieving the primary outcome (must have)
    • Phase 2: Enhances the primary outcome or enables secondary outcomes (should have)
    • Phase 3+: Additional value but not critical for initial outcomes (could have)
    • Out of Scope: Not aligned with prioritized outcomes or not feasible (won't have)
  • You MUST define clear acceptance criteria for Phase 1 that demonstrate outcome achievement
  • You MUST ensure Phase 1 represents 20-40% of total scope and effort (estimated roughly)
  • You MUST identify the minimum viable user journey or workflow
  • You MUST specify what will be explicitly excluded from Phase 1 and why
  • You MUST include detailed reasoning for all scoping decisions:
    • For each feature included in Phase 1 Core: Explain how it directly enables the top-priority outcome AND why it can't be deferred
    • For each feature excluded from Phase 1: Explain why it's not essential for the primary outcome OR why including it would exceed the 12-week constraint
    • For the 8-12 week timeline estimate: Explain the key factors that drove the estimate and how it fits within the 2-3 month hard constraint
    • If the timeline approaches or exceeds 12 weeks: Explain what additional descoping options exist to bring it under the constraint
    • For standalone value claim: Explain what value can be measured/delivered independently within the 2-3 month window
    • For risk assessment: Explain what makes the technical risk "manageable" or identify specific concerns
    • For feature categorization decisions: Explain the logic (e.g., "Categorized as Phase 2 because it enhances reporting but basic reporting in Phase 1 is sufficient to measure the outcome...")
  • You MUST save Phase 1 scope definition to "{working_directory}/phase1-mvp-scope-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "phase1-mvp-scope-draft.md" to "phase1-mvp-scope.md"
  • You MUST NOT include features solely because they're easy if they don't contribute to the prioritized outcome

Review Summary Requirements for phase1-mvp-scope.md:

  • Key Findings: % of total PRD in Phase 1, core feature count, estimated timeline (must be ≤12 weeks), primary user journey
  • Red Flags: Timeline exceeds 12 weeks (hard constraint violation), features that seem critical but were excluded, timeline feasibility concerns given team size, standalone value not clearly demonstrable, scope too large for 2-3 month window, dependencies on excluded features
  • Critical Decisions: Which features made the cut vs. excluded, timeline estimate rationale and how it stays within 12-week limit, how standalone value is achieved within the constraint
  • Critical Assumptions: User adoption assumptions, technical feasibility assumptions, dependency assumptions, team velocity assumptions
  • Approval Checklist: Verify Phase 1 directly achieves top-priority outcome, verify timeline is ≤12 weeks (2-3 months), verify excluded features aren't critical, verify standalone value is measurable, verify scope is focused enough to fit timeline constraint

Human Review Gate: After generating phase1-mvp-scope-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the core features included and major exclusions
  • Ask the user to review the draft artifact
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft artifact and re-present for approval
  • Once approved, rename "phase1-mvp-scope-draft.md" to "phase1-mvp-scope.md"
  • Only proceed to Step 2 after the artifact is approved and renamed

2. Structure Subsequent Milestone Phases

Organize remaining scope into logical milestone phases that progressively unlock value.

Constraints:

  • You MUST create 2-5 total milestone phases (including Phase 1 from Step 1) based on total project scope
  • You MUST structure each phase (Phase 2, 3, etc.) to:
    • Deliver standalone business value that can be measured and validated
    • Build logically on previous phases without requiring future phases to be valuable
    • Represent 2-8 weeks of effort (sweet spot for risk management) - estimated roughly
    • Have clear acceptance criteria and success metrics
    • Address a specific business outcome or user need from the SOP 1 outcomes list
  • You MUST sequence phases based on:
    • Technical dependencies: Foundational work before dependent features (e.g., can't do advanced analytics before basic data collection)
    • Business value priority: Higher value outcomes earlier when possible (reference SOP 1 prioritization scores)
    • Risk mitigation: Validate uncertain approaches early OR defer high-risk work until foundation is proven
    • Logical user experience progression: Natural evolution of functionality
  • You MUST balance phase sizing to avoid:
    • Phases too small (excessive overhead, fragmented delivery, < 2 weeks)
    • Phases too large (increased risk, delayed value, > 8 weeks)
  • You MUST calculate rough total scope for each phase (estimate based on feature complexity)
  • You MUST identify phase dependencies and prerequisites (e.g., "Phase 3 requires Phase 2 data pipeline completion")
  • You MUST define phase-specific success criteria and deliverables for each phase
  • You MUST include detailed reasoning for phase structure:
    • For each phase's scope: Explain what outcome it delivers and why that outcome was selected for that phase
    • For sequencing decisions: Explain why Phase 2 comes before Phase 3, etc. (dependencies, value, risk)
    • For phase sizing: Explain why the phase is the estimated size (what's included, complexity factors)
    • For phase dependencies: Explain the technical or logical dependency chain
    • For any phases that combine multiple outcomes: Explain why they were grouped together
    • For timeline balancing: Explain the trade-offs made in phase sizing
  • You SHOULD aim for relatively even phase sizes when possible to ensure steady value delivery
  • You MUST save phase structure to "{working_directory}/milestone-phase-structure-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "milestone-phase-structure-draft.md" to "milestone-phase-structure.md"

Review Summary Requirements for milestone-phase-structure.md:

  • Key Findings: Total phase count, phase sizes (weeks), outcomes addressed per phase, dependency chain summary
  • Red Flags: Phases with unclear value proposition, phases too large/small, dependency bottlenecks, phases that can't stand alone, uneven phase sizing creating risk concentration
  • Critical Decisions: Phase sequencing rationale, how outcomes were distributed across phases, dependency management approach
  • Critical Assumptions: Technical dependency assumptions, value delivery assumptions, effort estimates used for sizing
  • Approval Checklist: Verify each phase delivers standalone value, verify sequencing makes sense, verify phase sizes are balanced, verify no major dependencies missed, verify outcomes map to phases logically

Human Review Gate: After generating milestone-phase-structure-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the phase progression and key dependencies
  • Ask the user to review the draft artifact
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft artifact and re-present for approval
  • Once approved, rename "milestone-phase-structure-draft.md" to "milestone-phase-structure.md"
  • Only proceed to Step 3 after the artifact is approved and renamed

3. Systematic Task Decomposition for All Phases

Break down each phase's scope into granular, estimable task units using hierarchical decomposition informed by the technical context.

Constraints:

  • You MUST reference the technical context document when decomposing tasks:
    • Use specified technology stack to create technology-specific tasks (e.g., "React component development" vs. generic "UI development")
    • Include design/architecture tasks if technology decisions are TBD (e.g., "Select and justify database technology", "Design API architecture")
    • Include integration-specific tasks based on existing systems identified (e.g., "ERP API integration", "SSO configuration")
    • Include learning/training tasks if team lacks familiarity with required technologies
    • Account for compliance requirements in task breakdown (e.g., "HIPAA compliance audit", "Security review")
  • You MUST decompose tasks to a granularity where each unit represents 0.5-5 days of effort (or 4-40 hours)
  • You MUST use recursive decomposition for complex features:
    • Start with high-level feature or requirement
    • Break into logical sub-components
    • Continue breaking down until reaching estimable units
    • Document decomposition tree showing parent-child relationships
  • You MUST identify the following work categories for each decomposed task:
    • Core implementation work
    • Testing and quality assurance
    • Integration and configuration
    • Documentation and knowledge transfer
    • Code review and refinement
    • Debugging and troubleshooting buffer
    • Deployment and release activities
  • You MUST create a hierarchical task structure for each phase:
    • Phase → Feature Group → Feature → Implementation Tasks → Sub-tasks (if needed)
  • You MUST identify dependencies between tasks within and across phases
  • You MUST flag tasks that span multiple features or have cross-cutting concerns
  • You MUST ensure all PRD features categorized in Phase 1/2/3+ have corresponding task breakdowns
  • You MUST include detailed reasoning in task decomposition:
    • For each major feature decomposition: Explain the breakdown logic (e.g., "User Authentication broken into: OAuth integration (external dependency), Session management (core logic), Role-based access (authorization layer)")
    • For each task's work categories: Explain what testing/integration/documentation is needed and why
    • For task granularity: If a task is larger than 5 days, explain why it can't be decomposed further OR decompose it
    • For dependencies: Explain the technical or logical reason for the dependency
    • For cross-cutting tasks: Explain what features they impact and why they can't be isolated
    • For estimation guidance: Provide context on task complexity to inform later estimation
  • You SHOULD create a dependency graph showing critical path and parallelizable work
  • You MUST save the decomposition to "{working_directory}/task-decomposition-draft.md" with hierarchical structure
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "task-decomposition-draft.md" to "task-decomposition.md"

Review Summary Requirements for task-decomposition.md:

  • Key Findings: Total task count, tasks per phase, average task granularity, critical path length, parallelizable work percentage
  • Red Flags: Tasks too large (>5 days) or too small (<0.5 days), missing task categories (e.g., no testing tasks), dependency cycles, orphan tasks not mapped to features, work categories missing (documentation, testing), features without corresponding tasks
  • Critical Decisions: Decomposition approach used, where decomposition stopped, how dependencies were identified
  • Critical Assumptions: Granularity assumptions, dependency assumptions, work category assumptions (e.g., "assumed unit tests sufficient, integration tests not needed")
  • Approval Checklist: Verify all Phase 1/2/3+ features have task breakdowns, verify task granularity is estimable (0.5-5 days), verify dependencies make sense, verify no major work categories missing, verify critical path is identified

Human Review Gate: After generating task-decomposition-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the total task count and any concerning patterns (e.g., many large tasks)
  • Ask the user to review the draft artifact
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft artifact and re-present for approval
  • Once approved, rename "task-decomposition-draft.md" to "task-decomposition.md"
  • Only proceed to Step 4 after the artifact is approved and renamed

4. Task Classification and Novelty Assessment

Classify each task by novelty to inform estimation approach in SOP 3.

Constraints:

  • You MUST classify each major task or task group as one of:
    • Routine: Similar tasks completed before, well-understood patterns, proven approaches (e.g., "Standard CRUD operations for inventory items")
    • Moderate Complexity: Some novel aspects, mostly familiar territory (e.g., "OAuth integration - team has done SSO before but not Google OAuth specifically")
    • Novel/Innovative: Unprecedented work, new technologies or patterns, high uncertainty (e.g., "Real-time WebSocket sync - team hasn't built real-time features before")
  • You MUST provide classification rationale for each task or task group based on:
    • Team experience with similar technologies
    • Availability of proven patterns or libraries
    • Clarity of requirements
    • Technical uncertainty or unknowns
    • Integration complexity
  • You MUST identify tasks that may benefit from:
    • Technical spike or proof-of-concept before full implementation
    • RAG-enhanced few-shot examples (if historical data available)
    • Enhanced chain-of-thought reasoning during estimation
    • Ensemble estimation methods (multiple independent estimates)
  • You MUST flag "almost right" tasks - work that appears simple but commonly requires significant debugging or refinement:
    • AI-generated code that needs validation
    • Complex business logic with edge cases
    • Integration points with external systems
    • Security-critical implementations
  • You MUST calculate novelty distribution for the project:
    • % of tasks that are Routine
    • % of tasks that are Moderate
    • % of tasks that are Novel
  • You MUST include detailed reasoning for classifications:
    • For each Routine classification: Explain what similar work has been done or what proven patterns exist
    • For each Moderate classification: Explain the familiar aspects AND the novel aspects
    • For each Novel classification: Explain what makes it unprecedented and what unknowns exist
    • For "almost right" flags: Explain the specific risks (e.g., "OAuth callback handling has many edge cases that AI-generated code typically misses")
    • For spike recommendations: Explain what needs to be validated and why
  • You MUST save task classification to "{working_directory}/task-classification-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "task-classification-draft.md" to "task-classification.md"

Review Summary Requirements for task-classification.md:

  • Key Findings: Novelty distribution (% routine/moderate/novel), highest-risk novel tasks, "almost right" task count, spike recommendations
  • Red Flags: High % of novel tasks (>30%), novel tasks on critical path, novel tasks in Phase 1, "almost right" tasks without adequate buffers, classifications that seem incorrect given team experience
  • Critical Decisions: Classification criteria used, boundary cases (routine vs. moderate), spike recommendations
  • Critical Assumptions: Team experience assumptions, technology maturity assumptions, pattern availability assumptions
  • Approval Checklist: Verify classifications match team's actual experience, verify novel tasks are correctly identified, verify "almost right" risks are flagged, verify spike recommendations are reasonable, verify Phase 1 isn't overloaded with novel work

Human Review Gate: After generating task-classification-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the novelty distribution and any high-risk novel tasks
  • Call out spike recommendations
  • Ask the user to review the draft artifact and validate classifications based on their team's experience
  • Wait for explicit approval before proceeding
  • If feedback is provided (e.g., "our team has actually done WebSocket work before"), update the draft artifact and re-present for approval
  • Once approved, rename "task-classification-draft.md" to "task-classification.md"
  • Only proceed to Step 5 after the artifact is approved and renamed

5. Generate SOP 3 Execution Prompt

Create the complete execution prompt for the next SOP in the sequence, including all necessary parameters and file paths.

Constraints:

  • You MUST generate a prompt that includes:
    • Clear instruction to execute SOP 3 (Estimation & Risk Analysis)
    • All file paths to artifacts generated in SOP 1 and SOP 2
    • The original prd_document path
    • Any optional parameters that should carry forward
    • The working_directory path
  • You MUST format the prompt as a ready-to-use execution command
  • You MUST include a brief summary of what was accomplished in SOP 2 to provide context
  • You MUST highlight the milestone structure (number of phases, timeline range)
  • You MUST highlight the task decomposition (total task count, novelty distribution)
  • You SHOULD include any important caveats, assumptions, or notes that SOP 3 should be aware of

Prompt Template:

✅ SOP 2 (Milestone Scoping & Task Decomposition) completed successfully.

## Summary of SOP 2 Outputs:
- **Milestone Structure**: [X] phases spanning [Y-Z] weeks
- **Phase 1 Scope**: [Brief description of MVP/Phase 1]
- **Total Tasks**: [N] tasks decomposed
- **Novelty Distribution**: [X]% Routine, [Y]% Moderate, [Z]% Novel
- **Artifacts Generated**:
  - phase1-mvp-scope.md
  - milestone-phase-structure.md
  - task-decomposition.md
  - task-classification.md

## Next Step:
Execute **SOP 3: Estimation & Risk Analysis** to generate effort estimates and risk assessments for all tasks and phases.

---

## SOP 3 Execution Prompt:

Execute the Estimation & Risk Analysis SOP (proposalgen-sop-3-estimation-risk.sop.md) with the following parameters:

**Required Parameters:**
- prd_document: "[path/to/prd-source.md]"
- prd_breakdown: "[path/to/prd-breakdown.md]"
- business_outcomes: "[path/to/business-outcomes.md]"
- outcome_prioritization: "[path/to/outcome-prioritization.md]"
- phase1_mvp_scope: "[path/to/phase1-mvp-scope.md]"
- milestone_phase_structure: "[path/to/milestone-phase-structure.md]"
- task_decomposition: "[path/to/task-decomposition.md]"
- task_classification: "[path/to/task-classification.md]"

**Optional Parameters:**
[Include any optional parameters that were provided in SOP 1 or SOP 2]
- estimation_granularity: "[hours/days/story-points, default: days]"
- historical_data_source: "[path if available]"
- team_context: "[team composition, skill levels, experience]"
- timeline_constraints: "[value if provided]"
- budget_range: "[value if provided]"
- risk_tolerance: "[value, default: moderate]"
- working_directory: "[path/to/proposal-outputs]"

**Important Context for SOP 3:**
[Include any important notes, caveats, or assumptions that SOP 3 should be aware of]
- [Note 1: e.g., "Phase 1 has 3 novel tasks requiring technical spikes"]
- [Note 2: e.g., "ERP integration dependency flagged as high-risk"]
- [Note 3: e.g., "Total of 127 tasks across 4 phases"]

Output:

  • You MUST present this prompt to the user
  • You MUST clearly indicate that SOP 2 is complete and they should execute the prompt above when ready to proceed
  • You SHOULD remind the user that they can modify any of the parameters before executing SOP 3 if needed

Review Summary Template

Every artifact generated in this SOP MUST include a Review Summary section at the very top, using this exact template:

# [Artifact Name]

## 📋 REVIEW SUMMARY
**Status**: Ready for Review
**Generated**: [ISO 8601 timestamp]
**Confidence Level**: [X/100] - [High (80-100) / Medium (60-79) / Low (40-59) / Very Low (<40)]

### 🎯 Key Findings
- [3-5 bullet points with the most important conclusions]
- [Key numbers, decisions, or outcomes]
- [What changed from inputs to outputs]

### ⚠️ RED FLAGS & SCRUTINY AREAS

**High Priority Issues:**
- [Things that are risky, uncertain, or could derail the project]
- [Assumptions that might be wrong]
- [Gaps in information or analysis]

**Moderate Concerns:**
- [Areas that need validation but aren't show-stoppers]
- [Ambiguities that should be clarified]

### 🔑 Critical Decisions & Assumptions

**Decisions Made:**
- [Key choices that affect downstream work]
- [Trade-offs and rationale]

**Critical Assumptions:**
- [Assumptions this analysis depends on]
- [What happens if these assumptions are wrong]

### 📊 Quantitative Summary
- [Key metrics, counts, totals relevant to this artifact]
- [Percentages, scores, confidence intervals]

### ✅ Approval Checklist
Before approving, verify:
- [ ] [Specific thing #1 to check - tailored to this artifact]
- [ ] [Specific thing #2 to check - tailored to this artifact]
- [ ] [Specific thing #3 to check - tailored to this artifact]
- [ ] No major elements overlooked
- [ ] Assumptions are reasonable and documented

**Action**: If approved, proceed to next step. If issues found, document feedback and request revisions.

---

## DETAILED ANALYSIS
[Full artifact content follows...]

Examples

Example Input

prd_document: "./inventory-mgmt-proposal/prd-source.md"
prd_breakdown: "./inventory-mgmt-proposal/prd-breakdown.md"
business_outcomes: "./inventory-mgmt-proposal/business-outcomes.md"
business_context_analysis: "./inventory-mgmt-proposal/business-context-analysis.md"
outcome_prioritization: "./inventory-mgmt-proposal/outcome-prioritization.md"
top_priority_outcome: "Achieve accurate real-time inventory visibility to reduce stock-outs"
timeline_constraints: "CEO wants to see progress before Q4 budget cycle (6 months out). Year-end inventory requires system operational by December."
budget_range: "$180k-$250k, prefer phased investment to spread cost"
resource_constraints: "3 developers: 1 senior (5 years experience), 2 mid-level (2-3 years). Limited IT staff for support."
risk_tolerance: "conservative"
working_directory: "./inventory-mgmt-proposal"

Example Output Summary (phase1-mvp-scope.md excerpt)

# Phase 1 / MVP Scope Definition

## 📋 REVIEW SUMMARY
**Status**: Ready for Review
**Generated**: 2025-01-15T16:45:00Z
**Confidence Level**: 82/100 - High

### 🎯 Key Findings
- **Phase 1 Focus**: Achieve accurate real-time inventory visibility for top 100 SKUs to reduce stock-outs
- **Scope Size**: 4 core features, ~25% of total PRD scope
- **Timeline Estimate**: 6 weeks (conservative estimate given team composition)
- **Standalone Value**: Can measure stock-out reduction and inventory accuracy within 2 weeks of deployment
- **Feature Exclusions**: ERP integration, automated reordering, advanced analytics all deferred to later phases

### ⚠️ RED FLAGS & SCRUTINY AREAS

**High Priority Issues:**
- **ERP integration excluded from Phase 1**: This means initial data entry will be manual or imported via CSV. Creates risk of low adoption if data entry burden is too high. Mitigated by focusing on top 100 SKUs only.
- **Timeline assumes no major barcode scanning issues**: If barcode scanner integration proves difficult, could add 1-2 weeks. Recommend barcode vendor research before Phase 1 kickoff.
- **Standalone value requires operations team adoption**: If team doesn't use the system, we can't measure stock-out reduction. Change management critical.

**Moderate Concerns:**
- Limiting to top 100 SKUs may not show dramatic impact if stock-outs are spread across long tail of products
- Mobile app scope is minimal (scanning only) - may disappoint users expecting more functionality
- No offline mode in Phase 1 could limit warehouse usability if WiFi is spotty

### 🔑 Critical Decisions & Assumptions

**Decisions Made:**
- **Included barcode scanning in Phase 1** despite technical risk because manual inventory counts won't show enough improvement to validate outcome (Reasoning: Stock-out reduction requires real-time data capture, which requires scanning)
- **Excluded ERP integration from Phase 1** to reduce risk and complexity (Reasoning: ERP integration is highest-risk item from SOP 1; defer until proven value with standalone system)
- **Limited scope to top 100 SKUs** vs. all 2,000 SKUs (Reasoning: 80/20 rule - top 100 SKUs likely drive 70-80% of stock-outs; focus proves value faster with less data migration)
- **Deferred automated reordering to Phase 3** (Reasoning: Requires accurate historical data from Phases 1-2; can't automate what we can't measure yet)

**Critical Assumptions:**
- Operations team will adopt mobile scanning workflow (change management success assumed)
- Top 100 SKUs are responsible for majority of stock-outs (assumption based on typical inventory patterns, not validated with client data)
- 6 weeks is achievable with 3-person team @ 75% productivity (assumes no major blockers, mid-level devs are productive)
- Barcode scanner integration is straightforward (assumed standard barcode scanner hardware, not specialized equipment)
- Manual data entry for 100 SKUs is acceptable short-term burden

### 📊 Quantitative Summary
- **Features in Phase 1 Core**: 4 (Inventory tracking, Barcode scanning, Basic reporting, User auth)
- **Features in Phase 2**: 3 (ERP integration, Advanced reporting, Admin portal)
- **Features in Phase 3+**: 5 (Automated reordering, Predictive analytics, Mobile enhancements, etc.)
- **Features Out of Scope**: 2 (Multi-warehouse, External API)
- **% of Total PRD in Phase 1**: ~25%
- **Estimated Timeline**: 6 weeks
- **SKU Coverage**: 100 SKUs (top 5% by volume/value)

### ✅ Approval Checklist
Before approving, verify:
- [ ] Phase 1 scope directly achieves top-priority outcome (inventory visibility → stock-out reduction)
- [ ] 6-week timeline is realistic given team composition and conservative risk tolerance
- [ ] Top 100 SKU limitation is acceptable and will show meaningful results
- [ ] ERP integration exclusion is acceptable (manual/CSV data entry short-term)
- [ ] Barcode scanning inclusion makes sense despite technical risk
- [ ] Standalone value is measurable (can track stock-outs and inventory accuracy for 100 SKUs independently)
- [ ] No major elements overlooked
- [ ] Assumptions are reasonable and documented

**Action**: If approved, proceed to Step 2 (structure subsequent phases). If top 100 SKU limit is too restrictive or ERP integration is required in Phase 1, provide feedback.

---

## DETAILED ANALYSIS

### Top-Priority Outcome (from SOP 1)

**Outcome**: Achieve accurate real-time inventory visibility to reduce stock-outs
**Priority Score**: 4.35/5.00
**Key Success Metrics**:
- Reduce stock-outs by 50%
- Achieve inventory accuracy >95%
- Show measurable improvement within 12 weeks

### Phase 1 Core Features (Must Have)

#### 1. Web-Based Inventory Tracking System

**Inclusion Reasoning**: Core capability to achieve inventory visibility. Without a system to record and view inventory, we cannot measure accuracy or track stock levels. This is the foundational "must have" for the entire outcome.

**Scope**:
- Inventory item master data (top 100 SKUs: name, description, barcode, reorder point)
- Current stock levels by location
- Inventory movement history (receipts, issues, adjustments)
- Simple search and filter functionality

**Exclusions**:
- ❌ Multi-warehouse support (only single primary warehouse in Phase 1)
- ❌ Lot/batch tracking (defer to Phase 2 if needed)
- ❌ Serial number tracking (defer to Phase 2 if needed)

**Why Excluded**: Multi-warehouse, lot tracking, and serial numbers add complexity without contributing to core outcome of visibility for stock-out reduction. Business context mentions single warehouse operation.

#### 2. Barcode Scanning Mobile App (iOS/Android)

**Inclusion Reasoning**: Critical for real-time data capture. Manual data entry is too slow and error-prone to achieve 95% accuracy target. Operations team needs mobile scanning for receiving, picking, and cycle counting. This enables the "real-time" aspect of the outcome.

**Risk Note**: This is a Moderate-to-High complexity task (mobile development, barcode integration). Included despite risk because manual inventory counts won't show enough improvement to validate the outcome.

**Scope**:
- Receive inventory (scan barcode, enter quantity, location)
- Issue inventory (scan barcode, reduce quantity)
- Cycle count / inventory adjustment (scan barcode, enter actual count)
- Basic offline queue (queue scans if WiFi drops, sync when reconnected)

**Exclusions**:
- ❌ Full offline mode (defer to Phase 4)
- ❌ Advanced mobile features (photo capture, notes, etc.)
- ❌ Tablet optimization (phone-only initially)

**Why Excluded**: Full offline mode is complex and not essential for primary warehouse with WiFi. Advanced features add development time without directly improving inventory accuracy. Phone-first approach reduces testing burden.

#### 3. Basic Inventory Reports & Dashboard

**Inclusion Reasoning**: Must be able to measure the outcome. Reports showing current stock levels, stock-out incidents, and inventory accuracy are essential to validate that Phase 1 achieved the goal.

**Scope**:
- Current inventory levels (all 100 SKUs at a glance)
- Stock-out report (items at/below reorder point)
- Inventory accuracy report (cycle count variance)
- Inventory movement history (transaction log)
- Simple dashboard with key metrics (accuracy %, stock-out count, recent movements)

**Exclusions**:
- ❌ Advanced analytics (trend analysis, forecasting)
- ❌ Custom report builder
- ❌ Data export to Excel (defer to Phase 2)

**Why Excluded**: Advanced analytics and custom reporting are "nice to have" but basic reports are sufficient to measure the success metrics (50% stock-out reduction, 95% accuracy). Export can wait until Phase 2 when finance team needs it.

#### 4. User Authentication & Role-Based Access

**Inclusion Reasoning**: Security and accountability requirement. Must know who performed inventory transactions for audit trail and accuracy. Basic security is non-negotiable for any production system.

**Scope**:
- User login (username/password)
- Role-based access (Admin, Manager, Warehouse Staff)
- Basic permissions (view vs. edit inventory)
- Audit trail (who changed what, when)

**Exclusions**:
- ❌ SSO/Active Directory integration (defer to Phase 2)
- ❌ Advanced permissions (field-level, SKU-level)
- ❌ Password policies and MFA (defer to Phase 2 based on security review)

**Why Excluded**: SSO integration adds complexity and dependency on IT team (who has limited bandwidth per business context). Basic auth is sufficient for Phase 1 with small user base (operations team). Advanced permissions can wait until Phase 2 when broader rollout happens.

---

### Phase 1 Exclusions (and Rationale)

#### ERP Integration (Deferred to Phase 2)

**Why Excluded**:
- Highest-risk item identified in SOP 1 prioritization (Implementation Risk: 3/5 primarily due to ERP integration)
- Adds significant complexity and dependency on ERP team/vendor
- Business context mentions "limited IT staff" - ERP integration will require their support
- Phase 1 can prove value with standalone system before tackling complex integration
- Conservative risk tolerance suggests de-risking Phase 1 by deferring high-complexity items

**Short-Term Workaround**: Manual data entry or CSV import for initial 100 SKUs. With only 100 items, this is manageable.

**Phase 2 Addition**: Once Phase 1 proves inventory tracking works and team is adopted, tackle ERP integration for automated data sync.

#### Automated Reordering (Deferred to Phase 3)

**Why Excluded**:
- Ranked as separate outcome (#3 or #4 in SOP 1 prioritization) - not part of top-priority outcome
- Requires accurate historical data from Phase 1-2 to calculate reorder points effectively
- Can't automate decisions without first proving we can measure accurately
- Success criteria focuses on stock-out reduction through visibility, not automation

**Phase 3 Addition**: Once 2-3 months of accurate data collected in Phase 1-2, can build reordering algorithms on solid foundation.

#### Advanced Analytics & Reporting (Deferred to Phase 2-3)

**Why Excluded**:
- Not essential for core outcome - basic reports sufficient to measure 50% stock-out reduction and 95% accuracy
- Advanced analytics is more valuable once you have historical data (Phase 1 starts with zero history)
- Prioritization framework scored this lower on Time to Value

**Phase 2-3 Addition**: Add trend analysis, forecasting, and custom reports once baseline data exists.

#### All-SKU Coverage (Deferred to Phase 2)

**Why Excluded**:
- Focusing on top 100 SKUs (vs. all 2,000) reduces data migration effort by 95%
- 80/20 rule: Top 100 SKUs likely responsible for 70-80% of stock-outs and revenue
- Proves value faster with smaller scope
- Conservative risk tolerance suggests starting small

**Assumption**: Top 100 SKUs drive majority of stock-outs. This should be validated with client (e.g., "Which SKUs have most frequent stock-outs?")

**Phase 2 Addition**: Expand to all 2,000 SKUs once Phase 1 proves the system works for the critical items.

---

### Minimum Viable User Journey

**Scenario**: Operations Manager receives new shipment of a top-100 SKU

1. **Receive Inventory** (Mobile App)
   - Scan barcode on incoming pallet
   - App displays item name, confirms correct SKU
   - Enter quantity received
   - Confirm location (Warehouse A, Bin 23)
   - Submit → Inventory updated in real-time

2. **Check Stock Levels** (Web Dashboard)
   - Manager logs into web dashboard
   - Views current stock level for that SKU
   - Sees stock level increased by received quantity
   - Checks if item is still at/below reorder point

3. **Identify Stock-Outs** (Web Report)
   - Manager runs "Stock-Out Report"
   - Report shows all items at/below reorder point
   - Identifies items needing reorder
   - Takes action to reorder (manual process in Phase 1)

4. **Measure Accuracy** (Cycle Count)
   - Warehouse staff performs cycle count weekly
   - Scans item barcode, enters actual count
   - System calculates variance (expected vs. actual)
   - Manager reviews accuracy report
   - If variance exists, investigates and adjusts inventory

**Success Validation**: After 2 weeks of Phase 1 deployment, manager can compare:
- Stock-out incidents (before vs. after)
- Inventory accuracy % (cycle count variances)
- Time spent on manual inventory tracking (before: spreadsheets, after: mobile scanning)

---

### Acceptance Criteria for Phase 1

**Functional Criteria**:
- [ ] 100 SKUs loaded into system with correct barcode, reorder point data
- [ ] Mobile app can scan barcodes and update inventory (receive, issue, adjust)
- [ ] Web dashboard displays current inventory levels for all 100 SKUs
- [ ] Stock-out report identifies items at/below reorder point
- [ ] Inventory accuracy report shows cycle count variances
- [ ] User authentication works (login, role-based access)
- [ ] Audit trail captures all inventory transactions with user/timestamp

**Success Metrics** (measured 2 weeks post-deployment):
- [ ] Inventory accuracy >90% for top 100 SKUs (target: 95%, acceptable: 90%+ in Phase 1)
- [ ] Stock-out reduction >30% for tracked items (target: 50%, acceptable: 30%+ in Phase 1)
- [ ] User adoption >80% (80% of inventory transactions use the system vs. spreadsheets)
- [ ] Zero critical bugs affecting inventory accuracy

**Non-Functional Criteria**:
- [ ] Mobile app works on iOS and Android
- [ ] System response time <2 seconds for common operations
- [ ] 99% uptime during business hours (warehouse operations)
- [ ] Training completed for all operations team members
- [ ] Documentation delivered (user guide, admin guide)

---

### Timeline Estimate: 6 Weeks

**Constraint Compliance**: ✅ 6 weeks is well within the 12-week (2-3 month) hard constraint for Phase 1/MVP.

**Reasoning**:
- **Team composition**: 3 developers (1 senior, 2 mid-level) @ 75% productivity = ~11 person-weeks capacity
- **Conservative buffer**: Conservative risk tolerance requires higher buffer (25-30%)
- **Scope**: 4 core features, moderate complexity (mobile app + barcode integration)
- **Historical analogy**: Similar inventory tracking projects typically take 4-8 weeks for MVP (based on industry norms)
- **Constraint check**: Even with risk contingency (adding 2 weeks for barcode issues + dev productivity), total would be 8 weeks, still within 12-week limit

**Week-by-Week Breakdown** (rough):
- Week 1: Architecture, data model, development environment setup
- Week 2-3: Core inventory tracking (web) + user authentication
- Week 4: Mobile app development (barcode scanning)
- Week 5: Basic reporting + integration/testing
- Week 6: User acceptance testing, training, deployment

**Risks to Timeline**:
- Barcode scanner integration challenges could add 1 week (total: 7 weeks, still within constraint)
- Mid-level developers less productive than expected could add 1 week (total: 8 weeks, still within constraint)
- Scope creep (client requests additional features) could add 1-2 weeks (total: 8-9 weeks, approaching but within constraint)
- **Worst case with all risks**: ~9 weeks, which is within the 12-week hard limit

**Mitigation**:
- Clear scope freeze after approval of this Phase 1 definition. Any additional requests go to Phase 2.
- If timeline starts to exceed 10 weeks during execution, descope non-critical features to stay under 12-week constraint.

---

### Standalone Value Proposition

**Question**: Can Phase 1 be deployed and deliver value without Phase 2/3?

**Answer**: Yes. Phase 1 delivers standalone value:

1. **Measurable Outcome**: Can measure stock-out reduction and inventory accuracy for top 100 SKUs within 2 weeks of deployment
2. **Independent Operation**: System works without ERP integration (manual data entry acceptable for 100 SKUs)
3. **User Benefit**: Operations team gets mobile scanning and real-time visibility immediately
4. **ROI Validation**: Can calculate ROI based on stock-out reduction for top SKUs to justify Phase 2 investment

**What Happens if Phase 2 Doesn't Happen?**
- Client still has improved inventory visibility for their most critical 100 SKUs
- Client still achieves partial stock-out reduction (likely 30-40% vs. 50% target)
- Client still has foundation for future expansion (architecture, data model, user adoption)

**Go/No-Go Decision Point**: After Phase 1 acceptance, client can choose to:
-**Proceed to Phase 2** if results meet expectations (recommended)
- ⏸️ **Pause and iterate on Phase 1** if results are mixed (expand SKU coverage, refine features)
-**Stop after Phase 1** if results don't validate approach (unlikely but possible)

This optionality is the key benefit of milestone-based delivery for conservative risk tolerance clients.

Troubleshooting

PRD Scope Is Much Larger Than Expected

If the PRD contains 50+ features and defining an MVP seems impossible:

  • Focus ruthlessly on the single top-priority outcome from SOP 1
  • Apply the 80/20 rule: What 20% of features deliver 80% of the outcome value?
  • Consider whether the PRD is actually describing multiple products/phases
  • Don't try to fit everything into 3-4 phases - be willing to mark significant scope as "Future/Out of Scope"
  • Recommend a discovery phase if the scope is too unclear to confidently phase

Timeline Constraints Conflict with Milestone Sizing

If timeline requires phases smaller than 2 weeks or larger than 8 weeks:

  • DO NOT artificially compress phases to fit timeline - this creates risk
  • Present the realistic phase structure based on logical scope boundaries
  • Flag the timeline conflict in the Review Summary Red Flags
  • Provide options: (1) adjust timeline, (2) reduce scope, (3) add resources
  • Let the user and client decide how to resolve the constraint

Features Have Complex Dependencies Making Phasing Difficult

If many features depend on each other, making it hard to create standalone phases:

  • Identify the "core infrastructure" work that enables everything else
  • Consider a smaller Phase 0 or "Foundation Phase" focused on core architecture/infrastructure
  • Look for "vertical slices" - narrow end-to-end functionality rather than horizontal layers
  • Group tightly-coupled features into same phase even if it makes phase slightly larger
  • Document the dependency chain clearly so estimation in SOP 3 can account for it

Task Decomposition Hits Diminishing Returns

If you find yourself decomposing tasks into very small units (<0.5 days):

  • Stop decomposing - you've reached sufficient granularity
  • Tasks of 0.5-1 day are perfectly estimable
  • Remember: Estimation accuracy doesn't improve significantly below 0.5 day granularity
  • Focus decomposition effort on large/complex tasks (>5 days) rather than splitting small tasks further

Novelty Classification Is Unclear

If you can't determine whether a task is Routine vs. Moderate vs. Novel:

  • Ask clarifying questions about team experience with similar work
  • Default to the more conservative (higher novelty) classification when uncertain
  • Document the uncertainty in the reasoning
  • Flag for human review - the team knows their experience better than you do
  • Note that classification will be refined during SOP 3 estimation based on team input

Phase 1 Scope Seems Too Large or Too Small

If Phase 1 scope exceeds 12 weeks estimate or is less than 8 weeks:

  • Too Large (>12 weeks) - CRITICAL ISSUE:
    • This violates the hard 2-3 month constraint for MVP/Phase 1
    • You MUST ruthlessly descope to bring timeline under 12 weeks
    • Look for features that can be deferred to Phase 2 without losing core outcome
    • Consider narrowing the user base (e.g., top 100 SKUs vs. all SKUs, single department vs. company-wide)
    • Consider reducing feature depth (e.g., basic reporting vs. advanced analytics)
    • Challenge every "must have": Is this feature truly essential for the PRIMARY outcome?
    • If even minimal feature set exceeds 12 weeks, recommend:
      • Increasing team size to compress timeline
      • Selecting a different top-priority outcome that's achievable in 12 weeks
      • Discovery phase to refine scope before committing to Phase 1
    • DO NOT proceed with >12 week Phase 1 scope - this violates the fundamental constraint
  • Smaller Scope (8-10 weeks):
    • This is ideal - within the 2-3 month window with buffer
    • Verify standalone value is real and outcome is measurable
    • DO NOT add features just to reach 12 weeks - smaller focused scope is preferable
  • Too Small (<8 weeks):
    • Verify this isn't just "easy" work that doesn't deliver meaningful outcome
    • Ensure the outcome is substantial enough to validate the approach
    • Consider if additional features could enhance the same outcome without adding excessive risk
    • An 8-week Phase 1 that delivers measurable outcome is acceptable if genuine

User Wants All Work in Single Phase

If user/client resists milestone-based approach:

  • Explain the benefits: faster time-to-value, risk reduction, flexibility, fixed-price per milestone
  • Show timeline comparison: milestone approach typically delivers first value 8-12 weeks earlier
  • Highlight go/no-go flexibility after Phase 1
  • Consider hybrid: Larger Phase 1 (8-10 weeks) with smaller subsequent phases
  • If they insist on single phase, note that this becomes a waterfall fixed-price project with different risk profile (SOP 3-4 will need to adjust buffers significantly)

Professional Services Proposal - Estimation & Risk Analysis (SOP 3 of 4)

Overview

This is the third SOP in a 4-part series for creating fixed-price, milestone-based professional services project proposals. This SOP focuses on generating accurate effort estimates and comprehensive risk assessments for all tasks and milestone phases.

SOP Sequence:

  1. Discovery & Outcome Prioritization
  2. Milestone Scoping & Task Decomposition
  3. Estimation & Risk Analysis ← You are here
  4. Pricing & Proposal Generation

This SOP takes the task decomposition and classification from SOP 2 and generates detailed effort estimates using research-backed estimation techniques including chain-of-thought reasoning, three-point estimation, risk-adjusted buffers, and confidence scoring. The output provides the foundation for pricing in SOP 4.

Critical Research Insight: Direct LLM time estimates achieve only 16% accuracy. This SOP employs proven techniques including systematic task analysis, chain-of-thought reasoning, historical data retrieval, three-point estimation, and explicit risk buffers to achieve 70-85% accuracy for routine tasks and 55-65% for novel work.

AI Acceleration Methodology: This SOP applies configurable AI acceleration factors to reflect productivity gains from AI-assisted development. Default values (configurable via parameters):

  • Routine tasks: 40% reduction (ai_acceleration_routine = 0.40) - AI excels at boilerplate, CRUD, standard patterns
  • Moderate tasks: 25% reduction (ai_acceleration_moderate = 0.25) - AI helps but requires oversight
  • Novel tasks: 10% reduction (ai_acceleration_novel = 0.10) - Limited AI benefit due to lack of training data

Both Non-AI and AI-Adjusted estimates are documented, but only AI-Adjusted estimates are used for pricing and timeline calculations. AI-specific overhead (debugging, validation) is added separately to ensure quality.

Parameters

  • prd_document (required): Path to the original PRD source file.
  • prd_breakdown (required): Path to the structured PRD breakdown from SOP 1.
  • technical_context (required): Path to the technical context document from SOP 1 (informs risk assessment and team familiarity).
  • business_outcomes (required): Path to the business outcomes analysis from SOP 1.
  • outcome_prioritization (required): Path to the outcome prioritization from SOP 1.
  • phase1_mvp_scope (required): Path to the Phase 1/MVP scope definition from SOP 2.
  • milestone_phase_structure (required): Path to the milestone phase structure from SOP 2.
  • task_decomposition (required): Path to the detailed task decomposition from SOP 2.
  • task_classification (required): Path to the task novelty classification from SOP 2.
  • estimation_granularity (optional, default: "days"): Desired estimation unit - "hours", "days", or "story-points".
  • historical_data_source (optional): Path or description of historical project data for calibration and reference.
  • team_context (optional): Team composition, skill levels, experience with similar technologies, AI tool usage.
  • ai_acceleration_routine (optional, default: 0.40): AI acceleration factor for routine tasks (0.40 = 40% reduction). Range: 0.0-0.6.
  • ai_acceleration_moderate (optional, default: 0.25): AI acceleration factor for moderate complexity tasks (0.25 = 25% reduction). Range: 0.0-0.4.
  • ai_acceleration_novel (optional, default: 0.10): AI acceleration factor for novel/complex tasks (0.10 = 10% reduction). Range: 0.0-0.2.
  • timeline_constraints (optional): Hard deadlines or timing requirements.
  • budget_range (optional): Client's budget expectations or constraints.
  • risk_tolerance (optional, default: "moderate"): Organization's appetite for risk - "conservative", "moderate", or "aggressive".
  • working_directory (required): Directory path where all artifacts are saved (same as SOP 1 & 2).

Constraints for parameter acquisition:

  • You MUST ask for all required parameters upfront in a single prompt rather than one at a time
  • You MUST verify that all input files from SOP 1 and SOP 2 exist and are readable
  • You MUST validate AI acceleration parameters are within acceptable ranges:
    • ai_acceleration_routine: 0.0-0.6 (0-60% reduction), default 0.40
    • ai_acceleration_moderate: 0.0-0.4 (0-40% reduction), default 0.25
    • ai_acceleration_novel: 0.0-0.2 (0-20% reduction), default 0.10
    • If values exceed ranges, warn user and use maximum allowed value
    • If routine ≤ moderate ≤ novel (inverted), warn user - routine tasks should have highest acceleration
  • You MUST confirm successful acquisition of all parameters before proceeding
  • You SHOULD load and reference the SOP 2 artifacts (task decomposition, classification) to inform estimates

Steps

1. Historical Data Retrieval and Calibration

Retrieve and analyze similar historical tasks to inform estimates through calibration and reference class forecasting.

Constraints:

  • If historical_data_source is provided, You MUST retrieve and analyze historical data
  • You MUST attempt to find 5-10 similar historical tasks for each major task category or feature type
  • You MUST use similarity matching based on:
    • Task type and category (e.g., "Authentication", "API integration", "Mobile app development")
    • Technical complexity level
    • Technology stack and domain
    • Team experience level
  • You MUST extract actual vs. estimated times from historical data (if available)
  • You MUST calculate historical accuracy metrics:
    • Mean estimation error (actual / estimated)
    • Systematic biases (over-estimation vs. under-estimation patterns)
    • Variance by task type or novelty
  • You MUST identify calibration factors based on historical performance:
    • Example: "Backend API tasks historically take 1.3x initial estimates"
    • Example: "Mobile features with novel libraries take 1.8x estimates"
    • Example: "Integration tasks with third-party APIs take 1.5x estimates"
  • If historical data is NOT available, You MUST:
    • Clearly state this limitation in the artifact
    • Reduce overall confidence level by 15-25 points
    • Use industry-standard baselines and increase uncertainty buffers by 10-20%
    • Recommend establishing historical tracking for future estimates
  • You MUST include explicit reasoning for calibration:
    • For each calibration factor: Explain which historical tasks were analyzed and what patterns emerged
    • For missing calibration data: Explain what assumptions are being used instead
    • For industry baselines: Cite the source or explain the reasoning
  • You MUST save historical analysis to "{working_directory}/historical-calibration-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "historical-calibration-draft.md" to "historical-calibration.md"

Review Summary Requirements for historical-calibration.md:

  • Key Findings: Historical data availability, calibration factors identified, estimation accuracy patterns, task categories with most/least accurate estimates
  • Red Flags: No historical data available, historical accuracy <60%, high variance in estimates, systematic biases not accounted for, calibration factors seem unrealistic
  • Critical Decisions: Which historical tasks were used for calibration, how similarity was determined, which calibration factors to apply
  • Critical Assumptions: Historical data representativeness, team consistency assumptions, technology similarity assumptions
  • Approval Checklist: Verify calibration factors are reasonable, verify historical tasks are truly similar, verify biases are identified, verify industry baselines are credible if no historical data

Human Review Gate: After generating historical-calibration-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight key calibration factors that will be applied
  • Ask the user to validate calibration factors based on their experience
  • Wait for explicit approval before proceeding
  • If feedback is provided (e.g., "our team is actually faster at mobile dev"), update the draft and re-present
  • Once approved, rename to "historical-calibration.md"
  • Only proceed to Step 2 after approved and renamed

2. Risk and Complexity Assessment

Systematically identify risks, uncertainties, and complexity factors that affect effort for each task and phase, informed by the technical context.

Constraints:

  • You MUST reference the technical context document to inform risk assessment:
    • Use team familiarity assessments from technical context for Technical Novelty scoring
    • Use integration complexity assessments for Integration Complexity scoring
    • Use compliance requirements to elevate Quality Requirements scoring
    • Flag technology choices that are TBD as high uncertainty
    • Account for learning curve needs identified in technical context
  • You MUST rate each task from SOP 2 task decomposition on complexity dimensions (1-5 scale):
    • Technical Novelty (1-5): Familiarity with technology and patterns
      • 1: Completely novel, never done before, high learning curve
      • 3: Some novel aspects, mostly familiar patterns
      • 5: Routine work, done many times before
    • Integration Complexity (1-5): Number of systems and APIs involved
      • 1: Complex multi-system integration, many dependencies
      • 3: Moderate integration, 2-3 systems
      • 5: Standalone or simple integration, 1 system
    • Data Complexity (1-5): Schema design, migrations, data quality challenges
      • 1: Complex data model, migrations, quality issues
      • 3: Moderate data work, some migrations
      • 5: Simple data model, no migrations
    • Quality Requirements (1-5): Testing rigor, performance needs, compliance
      • 1: Extremely high quality bar (security-critical, compliance, high performance)
      • 3: Standard quality requirements (unit tests, basic performance)
      • 5: Low quality requirements (prototype, proof-of-concept)
    • Uncertainty Level (1-5): Specification clarity and requirement stability
      • 1: Very unclear requirements, likely to change significantly
      • 3: Mostly clear requirements, some ambiguity
      • 5: Crystal clear requirements, stable scope
  • You MUST identify specific risk factors for each task including:
    • Technical unknowns requiring research or proof-of-concept
    • Third-party dependencies with uncertain behavior (APIs, libraries, services)
    • Requirement ambiguity or likely scope changes
    • Technology maturity or team expertise gaps
    • "Almost right" risk - tasks likely to produce functional-but-not-quite-correct solutions requiring debugging
    • Hidden integration complexity with existing systems
    • Security or compliance risks
  • You MUST calculate a complexity multiplier for each task:
    • Average the 5 dimension scores to get overall complexity score (1-5 scale)
    • Apply multiplier based on overall score:
      • Low complexity (4.0-5.0 avg): 1.0-1.2x base estimate
      • Medium complexity (2.5-3.9 avg): 1.3-1.8x base estimate
      • High complexity (1.0-2.4 avg): 2.0-3.5x base estimate
  • You MUST identify project-level risks that affect multiple tasks or phases:
    • Technical risks (architecture decisions, technology selection, scalability)
    • Scope risks (requirement changes, scope creep, ambiguous features)
    • Resource risks (team availability, skill gaps, key person dependencies)
    • Client risks (delayed decisions, unavailable stakeholders, environment access delays)
    • External risks (third-party dependencies, regulatory changes, market shifts)
  • You MUST assess each project-level risk with:
    • Probability (1-5): How likely is this risk to occur? (1=unlikely, 5=very likely)
    • Impact (1-5): How severe if it occurs? (1=minor, 5=critical)
    • Severity = Probability × Impact (1-25 scale)
  • You MUST include detailed reasoning for all assessments:
    • For each task's complexity scores: Explain why each dimension received that score with specific evidence
    • For complexity multipliers: Explain the logic and how it compares to similar tasks
    • For "almost right" flags: Explain the specific debugging risks (e.g., "OAuth callback handling has many edge cases that AI-generated code typically misses")
    • For project-level risks: Explain probability and impact reasoning, cite evidence from PRD or business context
    • For risk severity calculations: Show the math and explain priority ordering
  • You MUST save risk assessment to "{working_directory}/risk-complexity-analysis-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "risk-complexity-analysis-draft.md" to "risk-complexity-analysis.md"

Review Summary Requirements for risk-complexity-analysis.md:

  • Key Findings: Average complexity score across all tasks, highest-risk tasks (complexity >3.0), project-level risk count, top 5 risks by severity
  • Red Flags: Many high-complexity tasks (>30%), high-complexity tasks on critical path, high-complexity tasks in Phase 1, project-level risks with severity >15, "almost right" tasks without adequate complexity multipliers
  • Critical Decisions: Complexity scoring rationale, multiplier calculations, which risks are most severe
  • Critical Assumptions: Team skill assumptions, technology maturity assumptions, requirement stability assumptions
  • Approval Checklist: Verify complexity scores match team's experience, verify high-risk tasks are correctly identified, verify project-level risks are comprehensive, verify "almost right" flags are reasonable

Human Review Gate: After generating risk-complexity-analysis-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the highest-complexity tasks and top project-level risks
  • Ask the user to validate complexity assessments
  • Wait for explicit approval before proceeding
  • If feedback is provided, update the draft and re-present
  • Once approved, rename to "risk-complexity-analysis.md"
  • Only proceed to Step 3 after approved and renamed

3. Base Effort Estimation with Chain-of-Thought

Generate initial effort estimates for each task using structured reasoning and explicit assumptions.

Constraints:

  • You MUST use chain-of-thought prompting for each estimate (do NOT provide direct estimates without reasoning)
  • You MUST structure reasoning for EACH task as follows:
    1. Task Description: What needs to be done
    2. Technical Approach: How it will be implemented (specific technologies, patterns, approaches)
    3. Implementation Steps: Break down the work into sequential steps
    4. Similar Historical Tasks: Reference similar tasks from Step 1 if available
    5. Complexity Factors: Note the key complexity drivers from Step 2
    6. Potential Challenges: Identify what could go wrong or take longer
    7. Base Estimate: Initial time estimate before adjustments
    8. Key Assumptions: List critical assumptions underlying the estimate
  • You MUST apply baseline estimates using task type patterns (adjust based on your estimation_granularity):
    • Simple CRUD operations: 0.5-1 day (routine) to 1-2 days (novel)
    • Complex business logic: 1-3 days (routine) to 4-6 days (novel)
    • Basic UI components: 0.5-1 day (simple) to 2-4 days (complex interactive)
    • Advanced UI/UX: 3-5 days (routine patterns) to 8-12 days (novel interactions)
    • API integrations: 1-3 days (well-documented) to 5-10 days (complex/poorly documented)
    • Database design: 0.5-1 day (simple entities) to 3-5 days (complex domains with migrations)
    • Authentication/authorization: 3-5 days (using framework) to 10-15 days (custom implementation)
    • Testing (unit + integration): 25-50% of implementation time
    • Documentation: 10-20% of implementation time
    • Code review and refinement: 15-25% of implementation time
  • You MUST adjust baselines using:
    • Task classification from SOP 2 (routine vs. moderate vs. novel)
    • Complexity multipliers from Step 2
    • Historical calibration factors from Step 1
  • You MUST apply the adjustments in sequence:
    1. Start with baseline estimate for task type
    2. Adjust for task novelty (routine = 1.0x, moderate = 1.3x, novel = 1.8x)
    3. Apply complexity multiplier from Step 2
    4. Apply historical calibration factor from Step 1 (if available)
    5. Result (Non-AI Estimate) = Base Estimate × Novelty × Complexity × Calibration
    6. Apply AI acceleration factor based on task novelty (using parameter values):
      • Routine tasks: ai_acceleration_routine (default 0.40 = 40% reduction, multiply by 1 - 0.40 = 0.60)
      • Moderate tasks: ai_acceleration_moderate (default 0.25 = 25% reduction, multiply by 1 - 0.25 = 0.75)
      • Novel/Complex tasks: ai_acceleration_novel (default 0.10 = 10% reduction, multiply by 1 - 0.10 = 0.90)
    7. Result (AI-Adjusted Estimate) = Non-AI Estimate × (1 - AI Acceleration Factor)
  • You MUST provide estimate in this format for EACH task:
    **Task**: [Task name]
    **Category**: [Task type, e.g., "API Integration"]
    **Novelty**: [Routine/Moderate/Novel]
    
    **Chain-of-Thought Reasoning**:
    - Technical Approach: [Description]
    - Implementation Steps: [1, 2, 3...]
    - Similar Tasks: [Historical reference if available]
    - Complexity Factors: [From Step 2 analysis]
    - Challenges: [What could go wrong]
    
    **Baseline Estimate**: [X days] (for [task type])
    **Novelty Adjustment**: [X days] × [1.0/1.3/1.8] = [Y days]
    **Complexity Multiplier**: [Y days] × [multiplier from Step 2] = [Z days]
    **Calibration Factor**: [Z days] × [calibration from Step 1] = [A days]
    **Non-AI Estimate**: [A] days
    **AI Acceleration**: [A days] × (1 - [ai_acceleration_routine/moderate/novel]) = [B days]
      (Example for routine task: A × (1 - 0.40) = A × 0.60 = B days, 40% reduction)
    **AI-Adjusted Estimate**: [B] days ← Used for pricing and totals
    
    **Key Assumptions**:
    - [Assumption 1]
    - [Assumption 2]
    - AI tools (GitHub Copilot, etc.) used effectively for [routine/moderate/novel] development work
    - AI acceleration factor: [ai_acceleration_routine/moderate/novel parameter value]
    
  • You MUST include explicit reasoning for every estimate showing how you arrived at the number
  • You MUST document BOTH Non-AI and AI-Adjusted estimates for each task
  • You MUST aggregate estimates by feature, phase, and total project using the AI-Adjusted estimates
  • You MUST clearly indicate which estimate is used for downstream calculations:
    • Non-AI Estimate: Documented for reference and transparency
    • AI-Adjusted Estimate: Used for all totals, pricing, and timeline calculations
  • You MUST include reasoning for AI acceleration factors:
    • Routine tasks (ai_acceleration_routine parameter, default 40% reduction): AI excels at boilerplate, CRUD operations, standard patterns
    • Moderate tasks (ai_acceleration_moderate parameter, default 25% reduction): AI helps with implementation but requires more human oversight
    • Novel tasks (ai_acceleration_novel parameter, default 10% reduction): AI provides limited benefit due to lack of training data and "almost right" risk
    • If non-default values are used, explain why they were adjusted (e.g., "Team has exceptional AI proficiency, increased routine acceleration to 50%")
  • You MUST save estimates to "{working_directory}/base-estimates-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "base-estimates-draft.md" to "base-estimates.md"
  • You MUST NOT provide single-point estimates without confidence intervals (Step 4 will add ranges)

Review Summary Requirements for base-estimates.md:

  • Key Findings: Total Non-AI estimate, Total AI-Adjusted estimate (used for pricing), AI acceleration %, estimate by phase (AI-Adjusted), largest task estimates, average task estimate
  • Red Flags: Tasks >5 days (AI-Adjusted) without further decomposition, AI-Adjusted estimates that seem too low/aggressive, missing reasoning, assumptions that seem wrong, calibration factors not applied, novelty adjustments not applied, AI acceleration factors not applied
  • Critical Decisions: Baseline selection logic, adjustment sequence, AI acceleration rationale, aggregation approach
  • Critical Assumptions: Team velocity assumptions, AI tool proficiency assumptions, technology learning curve assumptions, integration effort assumptions
  • Approval Checklist: Verify reasoning is clear and logical, verify adjustments are applied correctly, verify both Non-AI and AI-Adjusted estimates are documented, verify AI acceleration factors are appropriate, verify assumptions are documented, verify no tasks are missing estimates, verify aggregations use AI-Adjusted estimates

Human Review Gate: After generating base-estimates-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the total estimate and any outlier tasks
  • Ask the user to validate estimates based on their intuition and experience
  • Wait for explicit approval before proceeding
  • If feedback is provided (e.g., "this task seems underestimated"), update the draft with adjusted reasoning and re-present
  • Once approved, rename to "base-estimates.md"
  • Only proceed to Step 4 after approved and renamed

4. Three-Point Estimation (Optimistic/Likely/Pessimistic)

Generate three-scenario estimates to represent uncertainty and support risk management.

Constraints:

  • You MUST provide three estimates for each task:
    • Optimistic (Best Case): Assumes favorable conditions, minimal issues (20th percentile outcome)
    • Most Likely (Realistic): Expected case with normal challenges (50th percentile outcome) - this is the AI-Adjusted Estimate from Step 3
    • Pessimistic (Worst Case): Accounts for significant obstacles (80th percentile outcome)
  • You MUST use the AI-Adjusted Estimate as the baseline for three-point calculations (not the Non-AI estimate)
  • You MUST calculate three-point estimates using the task's novelty and uncertainty from SOP 2 and Step 2:
    • Routine tasks (low uncertainty):
      • Optimistic = Base × 0.8
      • Most Likely = Base × 1.0
      • Pessimistic = Base × 1.3
    • Moderate tasks (medium uncertainty):
      • Optimistic = Base × 0.75
      • Most Likely = Base × 1.0
      • Pessimistic = Base × 1.5
    • Novel tasks (high uncertainty):
      • Optimistic = Base × 0.7
      • Most Likely = Base × 1.0
      • Pessimistic = Base × 2.0
    • Very High Uncertainty tasks (unclear requirements, novel tech, complex integration):
      • Optimistic = Base × 0.7
      • Most Likely = Base × 1.0
      • Pessimistic = Base × 2.5
  • You MUST use PERT (Program Evaluation and Review Technique) to calculate expected value and standard deviation:
    • Expected Value = (Optimistic + 4×MostLikely + Pessimistic) / 6
    • Standard Deviation = (Pessimistic - Optimistic) / 6
    • Confidence Interval = Expected ± (1.96 × StdDev) for 95% confidence
  • You MUST aggregate three-point estimates at feature, phase, and project level:
    • Sum optimistic values for total optimistic
    • Sum most likely values for total most likely
    • Sum pessimistic values for total pessimistic
    • Calculate overall expected value using PERT formula
    • Calculate overall standard deviation = sqrt(sum of variances)
  • You MUST include explicit reasoning for range sizing:
    • For each task's pessimistic multiplier: Explain what scenarios drive the worst case (e.g., "Pessimistic assumes OAuth provider API has breaking changes requiring workaround")
    • For each task's optimistic multiplier: Explain what needs to go right (e.g., "Optimistic assumes barcode scanner integrates cleanly with no driver issues")
    • For uncertainty categorization: Explain why the task is high/medium/low uncertainty
  • You MUST present estimates in tabular format showing all three scenarios
  • You MUST save three-point estimates to "{working_directory}/three-point-estimates-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "three-point-estimates-draft.md" to "three-point-estimates.md"
  • You MUST NOT present only the "most likely" scenario - all three scenarios are required to communicate uncertainty

Review Summary Requirements for three-point-estimates.md:

  • Key Findings: Total estimates (optimistic/likely/pessimistic), PERT expected value, 95% confidence interval, widest uncertainty ranges
  • Red Flags: Pessimistic scenarios that seem unrealistic, narrow ranges that understate uncertainty, wide ranges indicating very high risk, total pessimistic exceeds budget significantly, Phase 1 pessimistic exceeds 12 weeks
  • Critical Decisions: Uncertainty categorization, range multipliers used, PERT calculation approach
  • Critical Assumptions: Probability distribution assumptions, independence of task uncertainties, risk correlation assumptions
  • Approval Checklist: Verify optimistic/pessimistic scenarios are realistic, verify ranges reflect true uncertainty, verify Phase 1 pessimistic ≤12 weeks, verify total pessimistic is within acceptable risk tolerance

Human Review Gate: After generating three-point-estimates-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the expected value and confidence interval for each phase and total project
  • Call out the widest uncertainty ranges and what drives them
  • CRITICAL: Flag if Phase 1 pessimistic estimate exceeds 12 weeks (violates hard constraint)
  • Ask the user to review and validate the ranges
  • Wait for explicit approval before proceeding
  • If Phase 1 pessimistic >12 weeks, recommend descoping Phase 1 before proceeding
  • If feedback is provided, update the draft and re-present
  • Once approved, rename to "three-point-estimates.md"
  • Only proceed to Step 5 after approved and renamed

5. Overhead, Buffers, and Hidden Costs

Add necessary overhead for project realities not captured in task-level estimates.

Constraints:

  • You MUST add the following overhead categories as percentages of core development effort:
    • Project coordination: 10-15% (meetings, status updates, planning, stakeholder communication)
    • Code review cycles: 15-20% (review time, addressing feedback, iterations)
    • Bug fixing and rework: 15-25% (defects discovered during development, not during testing)
    • Integration debugging: 10-20% (inter-component issues, deployment problems, environment issues)
    • Requirements clarification: 5-15% (ambiguity resolution, stakeholder discussions, scope clarification)
    • Context switching overhead: 5-10% (task interruptions, multi-tasking inefficiency)
  • If team_context indicates AI assistant usage, You MUST add AI-specific overhead:
    • "Almost right" debugging: 15-20% additional time for AI-generated code that needs refinement (especially for moderate/novel tasks)
    • AI suggestion validation: 8-12% time to verify correctness and security of AI outputs
    • Prompt engineering: 3-5% time to refine prompts and iterate to get useful AI assistance
    • NOTE: AI acceleration factors (Step 3) already account for productivity gains. These overhead percentages capture the additional validation/debugging work needed to ensure AI-generated code meets quality standards. Net effect is positive (acceleration > overhead) but both must be accounted for separately.
  • You MUST add uncertainty buffers based on risk analysis from Step 2:
    • Low risk components (complexity score 4.0-5.0): 15-25% buffer
    • Medium risk components (complexity score 2.5-3.9): 25-40% buffer
    • High risk components (complexity score 1.0-2.4): 40-60% buffer
    • Novel/innovative work: 60-100% buffer for unprecedented tasks
    • Very high uncertainty: 100-150% buffer for research-oriented work
  • You MUST adjust overhead for project characteristics:
    • Conservative risk tolerance: Use higher end of overhead ranges
    • Aggressive risk tolerance: Use lower end of overhead ranges
    • Fixed-price commitment: Add 10-20% contingency buffer
    • First-time client: Add 10-15% for relationship building and communication overhead
  • You MUST calculate overhead separately for optimistic, likely, and pessimistic scenarios:
    • Optimistic scenario: Minimum overhead percentages (assume smooth execution)
    • Most Likely scenario: Mid-range overhead percentages
    • Pessimistic scenario: Maximum overhead percentages (assume challenges occur)
  • You MUST present overhead as explicit line items, not hidden multipliers:
    Core Development: [X] days
    + Project Coordination (12%): [Y] days
    + Code Review (18%): [Z] days
    + Bug Fixing (20%): [A] days
    + Integration Debugging (15%): [B] days
    + Requirements Clarification (10%): [C] days
    + Context Switching (7%): [D] days
    + AI Debugging Overhead (25%): [E] days [if AI tools used]
    + Uncertainty Buffer (35%): [F] days
    = Total Effort: [Total] days
    
  • You MUST include detailed reasoning for overhead percentages:
    • For each overhead category: Explain why you chose the specific percentage within the range
    • For AI overhead: Explain the team's AI usage level and expected debugging needs
    • For uncertainty buffers: Explain how risk scores translate to buffer percentages
    • For project-specific adjustments: Explain the reasoning (e.g., "First-time client adds 15% due to need to establish communication patterns")
  • You MUST save overhead calculations to "{working_directory}/overhead-and-buffers-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "overhead-and-buffers-draft.md" to "overhead-and-buffers.md"
  • You MUST NOT omit buffers to make estimates appear more favorable - research shows systematic under-estimation when buffers are excluded

Review Summary Requirements for overhead-and-buffers.md:

  • Key Findings: Total overhead %, core effort vs. total effort, largest overhead categories, AI overhead (if applicable), uncertainty buffer %
  • Red Flags: Total overhead <30% (likely missing categories), total overhead >80% (may be over-buffered), AI overhead not included despite AI usage, uncertainty buffers don't match risk analysis, Phase 1 total exceeds 12 weeks
  • Critical Decisions: Overhead percentage selection rationale, buffer sizing logic, AI overhead inclusion/exclusion
  • Critical Assumptions: Team efficiency assumptions, AI productivity assumptions, risk buffer adequacy assumptions
  • Approval Checklist: Verify overhead categories are comprehensive, verify percentages match project risk profile, verify AI overhead is appropriate, verify Phase 1 total ≤12 weeks, verify buffers reflect risk analysis

Human Review Gate: After generating overhead-and-buffers-draft.md, you MUST:

  • Present the Review Summary to the user
  • Show the progression: Core → +Overhead → Total for each phase
  • CRITICAL: Flag if Phase 1 total (including all overhead) exceeds 12 weeks
  • Ask the user to validate overhead percentages
  • Wait for explicit approval before proceeding
  • If Phase 1 total >12 weeks, STOP and recommend descoping before proceeding
  • If feedback is provided, update the draft and re-present
  • Once approved, rename to "overhead-and-buffers.md"
  • Only proceed to Step 6 after approved and renamed

6. Final Estimation Report and Confidence Scoring

Compile all analysis into a comprehensive estimation report with confidence levels and validation guidance.

Constraints:

  • You MUST create a comprehensive assumptions document listing:
    • Technical assumptions: Frameworks, patterns, approaches, technology availability
    • Scope assumptions: What's included and excluded, feature depth, acceptance criteria
    • Team assumptions: Skill levels, availability, experience, AI tool proficiency
    • Process assumptions: Development methodology, review processes, deployment approach
    • Dependency assumptions: Third-party services, API availability, client-provided resources
    • Quality assumptions: Testing approach, performance requirements, acceptance criteria
    • Environmental assumptions: Tooling, infrastructure, access, development environments
  • You MUST identify assumption sensitivity - which assumptions have highest impact on estimates:
    • For each high-impact assumption: Explain what happens if the assumption is wrong
    • Estimate the impact magnitude (e.g., "If ERP API is poorly documented, could add 10-15 days to Phase 2")
  • You MUST calculate overall confidence score (0-100) for the estimate using:
    • Task novelty (max 30 points):
      • Routine tasks >70%: +30 points
      • Routine tasks 50-70%: +20 points
      • Routine tasks <50%: +10 points
    • Historical data availability (max 25 points):
      • Extensive historical data: +25 points
      • Some historical data: +15 points
      • No historical data: +5 points
    • Requirement clarity (max 20 points):
      • Crystal clear requirements: +20 points
      • Clear requirements with minor ambiguity: +15 points
      • Ambiguous or changing requirements: +5 points
    • Team experience (max 15 points):
      • Expert team with domain experience: +15 points
      • Experienced team: +10 points
      • Team learning new technologies: +5 points
    • Technical maturity (max 10 points):
      • Proven, stable technologies: +10 points
      • Established technologies: +7 points
      • Emerging or cutting-edge technologies: +3 points
  • You MUST provide confidence interpretation:
    • 80-100: High confidence, estimates suitable for fixed-price commitments
    • 60-79: Moderate confidence, suitable for planning with contingency
    • 40-59: Low confidence, require validation and may need refinement or discovery phase
    • Below 40: Very low confidence, recommend discovery phase or spike work before commitment
  • You MUST create a validation checklist for human reviewers:
    • Decomposition completeness - are all necessary tasks identified?
    • Estimate reasonableness - do estimates align with intuition and experience?
    • Risk coverage - are all significant risks accounted for in buffers?
    • Assumption validity - are assumptions realistic and documented?
    • Buffer adequacy - are buffers sufficient for project uncertainty?
    • Historical consistency - do estimates align with past performance?
    • Missing work - any categories of work overlooked (testing, documentation, deployment)?
    • Phase 1 constraint - does Phase 1 total remain ≤12 weeks in all scenarios?
  • You MUST create a final estimation report with these sections:
    • Executive Summary: Total effort (optimistic/likely/pessimistic), confidence level, Phase 1 estimate, critical risks (1 page max)
    • Estimation Approach: Methodology used, why it was selected, research basis
    • Task Classification Summary: Routine vs. moderate vs. novel breakdown
    • Effort Estimates by Phase: Three-point estimates for each milestone phase with PERT expected values
    • Total Effort Summary: Aggregated estimates with confidence intervals
    • Top Risk Factors: Top 10 risks affecting estimates from Step 2
    • Assumptions and Sensitivities: Critical assumptions and impact if wrong
    • Confidence Analysis: Confidence score breakdown and interpretation
    • Validation Notes: Areas requiring expert review or further investigation
    • Recommendations: Next steps (discovery phase, technical spikes, estimate refinement triggers)
  • You MUST present effort in multiple formats:
    • Total person-hours or person-days (both Non-AI and AI-Adjusted)
    • AI acceleration impact (% reduction from Non-AI to AI-Adjusted)
    • Breakdown by work category (development, testing, overhead, buffers)
    • Breakdown by phase (AI-Adjusted estimates)
    • Timeline estimate if team size is known (using AI-Adjusted days / team size / productivity factor)
  • You MUST include a visual summary table showing AI-Adjusted estimates (used for pricing):
    | Phase | Optimistic | Most Likely (AI-Adj) | Pessimistic | PERT Expected | Std Dev | 95% CI | Confidence |
    |-------|-----------|---------------------|-------------|---------------|---------|--------|------------|
    | Phase 1 | X days | Y days | Z days | E days | S days | E±C | 75% |
    | Phase 2 | X days | Y days | Z days | E days | S days | E±C | 70% |
    | TOTAL | X days | Y days | Z days | E days | S days | E±C | 72% |
    
  • You MUST include a comparison table showing AI impact:
    | Metric | Non-AI Estimate | AI-Adjusted Estimate | Reduction | Notes |
    |--------|----------------|---------------------|-----------|-------|
    | Total Effort | X days | Y days | Z% | Used for pricing |
    | Phase 1 | A days | B days | C% | Within 12-week constraint |
    
  • You MUST include sensitivity analysis showing how estimates change if key assumptions are wrong
  • You MUST include calibration guidance: "These estimates should be refined after [milestone] based on actual velocity"
  • You MUST include detailed reasoning throughout the report:
    • For confidence score: Explain how each component (novelty, historical data, clarity, team, tech) was assessed
    • For top risks: Explain why they're the most impactful
    • For assumptions: Explain why they're critical and what validates them
    • For recommendations: Explain the reasoning (e.g., "Recommend 3-day technical spike on WebSocket integration because team has no experience and pessimistic estimate is 2.5x optimistic")
  • You MUST save the final report to "{working_directory}/loe-estimate-draft.md"
  • You MUST include a Review Summary at the top of the artifact
  • After human approval, You MUST rename the file from "loe-estimate-draft.md" to "loe-estimate.md"
  • You MUST NOT present estimates as fixed commitments; emphasize they are probabilistic forecasts subject to refinement

Review Summary Requirements for loe-estimate.md:

  • Key Findings: Total PERT expected effort (AI-Adjusted), Non-AI vs AI-Adjusted comparison, AI acceleration %, Phase 1 PERT expected (AI-Adjusted), confidence score, top 3 risks, constraint compliance (Phase 1 ≤12 weeks)
  • Red Flags: Low confidence score (<60), Phase 1 exceeds 12 weeks in likely or pessimistic scenario (AI-Adjusted), AI acceleration seems too aggressive, very wide confidence intervals, high-impact assumptions not validated, top risks without mitigation plans
  • Critical Decisions: Final effort recommendations, AI acceleration rationale, confidence level interpretation, spike/discovery recommendations
  • Critical Assumptions: All high-impact assumptions from the assumptions document, AI tool proficiency assumptions
  • Approval Checklist: Verify total AI-Adjusted estimate is reasonable, verify AI acceleration is appropriate, verify both Non-AI and AI-Adjusted estimates are documented, verify Phase 1 ≤12 weeks (AI-Adjusted), verify confidence score reflects true uncertainty, verify all risks are captured, verify assumptions are documented, verify recommendations are actionable

Human Review Gate: After generating loe-estimate-draft.md, you MUST:

  • Present the Review Summary to the user
  • Highlight the total PERT expected value and confidence score
  • CRITICAL: Confirm Phase 1 remains ≤12 weeks in all scenarios (optimistic/likely/pessimistic)
  • Highlight top risks and recommendations
  • Ask the user to perform final validation
  • Wait for explicit approval before proceeding to Step 7
  • If feedback is provided, update the draft and re-present
  • Once approved, rename to "loe-estimate.md"
  • Only proceed to Step 7 after approved and renamed

7. Generate SOP 4 Execution Prompt

Create the complete execution prompt for the final SOP in the sequence.

Constraints:

  • You MUST generate a prompt that includes:
    • Clear instruction to execute SOP 4 (Pricing & Proposal Generation)
    • All file paths to artifacts generated in SOP 1, 2, and 3
    • The original prd_document path
    • Any parameters needed for pricing (rate cards, pricing model, margin targets)
    • Any optional parameters that should carry forward
    • The working_directory path
  • You MUST format the prompt as a ready-to-use execution command
  • You MUST include a brief summary of what was accomplished in SOP 3
  • You MUST highlight the total effort estimate and confidence level
  • You MUST highlight any Phase 1 constraint concerns if they exist
  • You SHOULD include any important caveats, assumptions, or notes that SOP 4 should be aware of

Prompt Template:

✅ SOP 3 (Estimation & Risk Analysis) completed successfully.

## Summary of SOP 3 Outputs:
- **Total Effort Estimate (AI-Adjusted)**: [X-Y] days (PERT Expected: [Z] days ± [CI])
- **Non-AI Estimate**: [N] days (for reference)
- **AI Acceleration Applied**: [%] reduction (Routine: [ai_acceleration_routine]%, Moderate: [ai_acceleration_moderate]%, Novel: [ai_acceleration_novel]%)
- **Phase 1 Estimate (AI-Adjusted)**: [A-B] days (PERT Expected: [C] days ± [CI])
- **Confidence Level**: [X/100] - [High/Medium/Low]
- **Top 3 Risks**: [Risk 1], [Risk 2], [Risk 3]
- **Phase 1 Constraint Compliance**: [✅ Phase 1 ≤12 weeks / ⚠️ Phase 1 approaches 12 weeks / ❌ Phase 1 exceeds 12 weeks]
- **Artifacts Generated**:
  - historical-calibration.md
  - risk-complexity-analysis.md
  - base-estimates.md (includes both Non-AI and AI-Adjusted)
  - three-point-estimates.md (AI-Adjusted)
  - overhead-and-buffers.md
  - loe-estimate.md (AI-Adjusted used for pricing)

## Next Step:
Execute **SOP 4: Pricing & Proposal Generation** to convert effort estimates into pricing, structure payment terms, and generate the client-facing proposal.

---

## SOP 4 Execution Prompt:

Execute the Pricing & Proposal Generation SOP (proposalgen-sop-4-pricing-proposal.sop.md) with the following parameters:

**Required Parameters:**
- prd_document: "[path/to/prd-source.md]"
- prd_breakdown: "[path/to/prd-breakdown.md]"
- business_outcomes: "[path/to/business-outcomes.md]"
- outcome_prioritization: "[path/to/outcome-prioritization.md]"
- business_context_analysis: "[path/to/business-context-analysis.md]"
- phase1_mvp_scope: "[path/to/phase1-mvp-scope.md]"
- milestone_phase_structure: "[path/to/milestone-phase-structure.md]"
- task_decomposition: "[path/to/task-decomposition.md]"
- task_classification: "[path/to/task-classification.md]"
- historical_calibration: "[path/to/historical-calibration.md]"
- risk_complexity_analysis: "[path/to/risk-complexity-analysis.md]"
- base_estimates: "[path/to/base-estimates.md]"
- three_point_estimates: "[path/to/three-point-estimates.md]"
- overhead_and_buffers: "[path/to/overhead-and-buffers.md]"
- loe_estimate: "[path/to/loe-estimate.md]"

**Optional Parameters:**
[Include any optional parameters that were provided in SOP 1-3]
- hourly_rate: "[your standard loaded rate, e.g., $150/hour]"
- pricing_model: "[fixed-price / time-and-materials / hybrid]"
- margin_target: "[desired margin %, e.g., 30%]"
- payment_terms: "[e.g., net-30, upon milestone acceptance]"
- timeline_constraints: "[value if provided]"
- budget_range: "[value if provided]"
- risk_tolerance: "[value, default: moderate]"
- working_directory: "[path/to/proposal-outputs]"

**Optional Parameters (if non-default AI acceleration used):**
- ai_acceleration_routine: "[value if non-default, e.g., 0.50]"
- ai_acceleration_moderate: "[value if non-default, e.g., 0.30]"
- ai_acceleration_novel: "[value if non-default, e.g., 0.15]"

**Important Context for SOP 4:**
[Include any important notes, caveats, or assumptions that SOP 4 should be aware of]
- All estimates are AI-Adjusted (Routine: [ai_acceleration_routine]%, Moderate: [ai_acceleration_moderate]%, Novel: [ai_acceleration_novel]%)
- AI acceleration assumes team proficiency with AI development tools (GitHub Copilot, etc.)
- [If non-default values used: Explain why, e.g., "Team has 2+ years AI tool experience, increased acceleration factors"]
- [Note 1: e.g., "Phase 1 pessimistic scenario is 11 weeks - very close to 12-week constraint"]
- [Note 2: e.g., "High uncertainty in ERP integration (Phase 2) - recommend 50% buffer in pricing"]
- [Note 3: e.g., "Confidence level is 72% - suitable for fixed-price with adequate buffers"]

Output:

  • You MUST present this prompt to the user
  • You MUST clearly indicate that SOP 3 is complete and they should execute the prompt above when ready to proceed
  • You SHOULD remind the user that they can modify any parameters before executing SOP 4 if needed
  • If Phase 1 exceeded the 12-week constraint, You MUST warn the user to address this before proceeding to SOP 4

Review Summary Template

Every artifact generated in this SOP MUST include a Review Summary section at the very top, using this exact template:

# [Artifact Name]

## 📋 REVIEW SUMMARY
**Status**: Ready for Review
**Generated**: [ISO 8601 timestamp]
**Confidence Level**: [X/100] - [High (80-100) / Medium (60-79) / Low (40-59) / Very Low (<40)]

### 🎯 Key Findings
- [3-5 bullet points with the most important conclusions]
- [Key numbers, decisions, or outcomes]
- [What changed from inputs to outputs]

### ⚠️ RED FLAGS & SCRUTINY AREAS

**High Priority Issues:**
- [Things that are risky, uncertain, or could derail the project]
- [Assumptions that might be wrong]
- [Gaps in information or analysis]

**Moderate Concerns:**
- [Areas that need validation but aren't show-stoppers]
- [Ambiguities that should be clarified]

### 🔑 Critical Decisions & Assumptions

**Decisions Made:**
- [Key choices that affect downstream work]
- [Trade-offs and rationale]

**Critical Assumptions:**
- [Assumptions this analysis depends on]
- [What happens if these assumptions are wrong]

### 📊 Quantitative Summary
- [Key metrics, counts, totals relevant to this artifact]
- [Percentages, scores, confidence intervals]

### ✅ Approval Checklist
Before approving, verify:
- [ ] [Specific thing #1 to check - tailored to this artifact]
- [ ] [Specific thing #2 to check - tailored to this artifact]
- [ ] [Specific thing #3 to check - tailored to this artifact]
- [ ] No major elements overlooked
- [ ] Assumptions are reasonable and documented

**Action**: If approved, proceed to next step. If issues found, document feedback and request revisions.

---

## DETAILED ANALYSIS
[Full artifact content follows...]

Examples

Example Input

[All SOP 1 and SOP 2 artifact paths]
estimation_granularity: "days"
historical_data_source: "project-history-database.json (contains 150 completed tasks from similar inventory/manufacturing projects)"
team_context: "3 developers: 1 senior (5 years experience with inventory systems), 2 mid-level (2-3 years). Team uses GitHub Copilot moderately. No prior barcode scanning experience."
timeline_constraints: "CEO wants to see progress before Q4 budget cycle (6 months out)."
budget_range: "$180k-$250k, prefer phased investment"
risk_tolerance: "conservative"
working_directory: "./inventory-mgmt-proposal"

Example Output Summary (loe-estimate.md excerpt)

# Level of Effort Estimation Report

## 📋 REVIEW SUMMARY
**Status**: Ready for Review
**Generated**: 2025-01-15T18:30:00Z
**Confidence Level**: 74/100 - Medium-High

### 🎯 Key Findings
- **Total Project Effort**: 285-510 days (PERT Expected: 368 days ± 42 days)
- **Phase 1 Effort**: 44-72 days (PERT Expected: 56 days ± 5 days)
- **Phase 1 Timeline**: 6.5-10.5 weeks with 3-person team (Expected: 8.2 weeks) ✅ Within 12-week constraint
- **Confidence**: 74/100 (Medium-High) - Suitable for fixed-price with adequate buffers
- **Top Risk**: ERP integration complexity in Phase 2 could add 15-25 days if API is poorly documented

### ⚠️ RED FLAGS & SCRUTINY AREAS

**High Priority Issues:**
- **Phase 1 pessimistic timeline is 10.5 weeks** - very close to 12-week hard constraint with minimal buffer
- **Barcode scanning has no team experience** - classified as Novel, 2.0x pessimistic multiplier applied, but could be higher if hardware issues arise
- **ERP integration in Phase 2 has 60% uncertainty buffer** - very high risk, recommend technical spike before committing to Phase 2 pricing
- **No historical data on mobile development** - mid-level devs learning React Native could take 1.5x longer than estimated

**Moderate Concerns:**
- AI debugging overhead (25%) may be conservative if team's Copilot proficiency is low
- Assumption that top 100 SKUs is sufficient to measure outcome - if not, Phase 1 scope may need expansion
- Testing overhead (35% of dev time) assumes standard quality bar - if client wants higher quality, could increase 10-15%

### 🔑 Critical Decisions & Assumptions

**Decisions Made:**
- Applied 1.3x calibration factor to backend tasks based on historical data showing team's backend estimates are optimistic
- Applied 1.8x novelty multiplier to barcode scanning due to zero team experience
- Used conservative (higher) overhead percentages throughout due to conservative risk tolerance
- Included 25% AI debugging overhead despite moderate Copilot usage because inventory domain is unfamiliar
- Applied 40% uncertainty buffer to ERP integration based on high complexity score (2.1/5.0) and third-party API dependency

**Critical Assumptions:**
- **Team velocity**: 3 developers @ 75% productivity (assumes some meetings, context switching, interruptions)
- **Top 100 SKU sufficiency**: Assumes focusing on top 100 SKUs is adequate to measure 50% stock-out reduction
- **Barcode scanner availability**: Assumes client has or will procure compatible barcode scanners (standard USB/Bluetooth models)
- **ERP API access**: Assumes client provides ERP sandbox environment and API documentation within 2 weeks of Phase 2 kickoff
- **Mid-level developer React Native learning**: Assumes 2-week learning curve included in estimates, developers can become productive
- **No major scope changes**: Assumes requirements are stable after Phase 1 scope freeze

### 📊 Quantitative Summary
- **Total Tasks**: 127 tasks across 4 phases
- **Task Novelty**: 45% Routine, 40% Moderate, 15% Novel
- **Average Complexity Score**: 3.2/5.0 (Medium complexity)
- **Total Overhead**: 42% of core development effort
- **Confidence Score**: 74/100 (Task Novelty: +20, Historical Data: +15, Requirement Clarity: +15, Team Experience: +10, Tech Maturity: +7, AI Usage: +7)
- **Phase 1 Constraint**: ✅ Pessimistic (72 days / 3 devs = 10.5 weeks) is within 12-week limit

### ✅ Approval Checklist
Before approving, verify:
- [ ] Total effort (368 days expected) seems reasonable for the scope
- [ ] Phase 1 estimate (56 days expected, 8.2 weeks) is achievable and ≤12 weeks in all scenarios
- [ ] Confidence score (74%) reflects true uncertainty
- [ ] Top risks (ERP integration, barcode scanning, mobile dev) are correctly identified and buffered
- [ ] Assumptions (team velocity, SKU sufficiency, API access) are realistic
- [ ] Overhead percentages (42% total) are appropriate for conservative risk tolerance
- [ ] Calibration factors (1.3x backend, 1.8x novel tasks) are validated by team's historical performance
- [ ] No major elements overlooked
- [ ] Assumptions are reasonable and documented

**Action**: If approved, proceed to SOP 4 for pricing. If Phase 1 timeline is too close to 12-week limit, consider descoping.

---

## DETAILED ANALYSIS

### Executive Summary

**Project**: Inventory Management System - Fixed-Price Milestone Delivery
**Total Effort**: 285-510 person-days (PERT Expected: 368 days ± 42 days, 95% CI: 284-452 days)
**Phase 1 Effort**: 44-72 person-days (PERT Expected: 56 days ± 5 days)
**Timeline**: Phase 1 = 8.2 weeks (3 devs @ 75%), Total Project = 54 weeks (3 devs @ 75%)
**Confidence**: 74/100 (Medium-High) - Suitable for fixed-price commitment with buffers
**Constraint Compliance**: ✅ Phase 1 within 12-week limit (pessimistic: 10.5 weeks)

**Top 10 Risks Affecting Estimates**:
1. **ERP Integration Complexity (Phase 2)**: Third-party API poorly documented or requires complex workarounds (+15-25 days)
2. **Barcode Scanner Hardware Issues (Phase 1)**: Incompatible hardware, driver problems, Bluetooth connectivity (+5-10 days)
3. **Mobile Development Learning Curve (Phase 1)**: Mid-level devs learning React Native slower than expected (+8-12 days)
4. **Requirement Ambiguity in Reporting (Phase 1-2)**: Report requirements not fully specified, rework needed (+5-8 days)
5. **Data Migration Complexity (Phase 2)**: Spreadsheet data quality worse than expected, cleaning required (+6-10 days)
6. **Client Decision Delays**: Delayed feedback, unavailable stakeholders (+10-15 days spread across phases)
7. **Scope Creep in Phase 1**: Client requests additional features during development (+5-10 days)
8. **AI Code Debugging Overhead Higher Than Expected**: Copilot generates incorrect business logic (+8-15 days)
9. **Third-Party Library Issues**: Barcode scanning library bugs or incompatibilities (+3-7 days)
10. **Integration Testing Complexity**: Inter-component issues more complex than expected (+5-8 days)

[... detailed estimates continue ...]

## Effort Estimates by Phase

| Phase | Optimistic | Most Likely | Pessimistic | PERT Expected | Std Dev | 95% CI | Timeline (3 devs @ 75%) | Confidence |
|-------|-----------|-------------|-------------|---------------|---------|--------|------------------------|------------|
| Phase 1: MVP | 44 days | 56 days | 72 days | 57 days | 4.7 days | 48-66 days | 8.2 weeks | 78% |
| Phase 2: ERP Integration | 80 days | 110 days | 165 days | 115 days | 14.2 days | 87-143 days | 16.7 weeks | 70% |
| Phase 3: Automation | 72 days | 98 days | 138 days | 101 days | 11.0 days | 79-123 days | 14.7 weeks | 72% |
| Phase 4: Enhancements | 89 days | 104 days | 135 days | 108 days | 7.7 days | 93-123 days | 15.7 weeks | 75% |
| **TOTAL** | **285 days** | **368 days** | **510 days** | **381 days** | **21.2 days** | **339-423 days** | **55.3 weeks** | **74%** |

**Phase 1 Constraint Check**:
- Optimistic: 44 days / 3 devs / 0.75 = 6.5 weeks ✅
- Most Likely: 56 days / 3 devs / 0.75 = 8.2 weeks ✅
- Pessimistic: 72 days / 3 devs / 0.75 = 10.5 weeks ✅ (within 12-week limit, but tight)
- **Status**: Phase 1 complies with 12-week constraint in all scenarios

[... detailed breakdown continues ...]

Troubleshooting

No Historical Data Available

If no historical data exists:

  • Clearly state in historical-calibration.md that confidence is reduced by 20-25 points
  • Use industry-standard baselines from research and best practices
  • Increase uncertainty buffers by 15-20 percentage points across all tasks
  • Apply wider pessimistic multipliers (2.0x → 2.5x for novel tasks)
  • Recommend shadow-mode estimation: track actuals vs. estimates to build historical database
  • Use more conservative overhead percentages (use high end of ranges)
  • Flag in every artifact Review Summary that lack of historical data increases uncertainty

Estimates Significantly Exceed Budget Constraints

If total estimate greatly exceeds stated budget:

  • DO NOT artificially reduce estimates to fit budget
  • Present estimates as-is based on analysis
  • In final report, clearly identify the gap between estimate and budget
  • Provide options in recommendations section:
    • Reduce scope (descope phases, reduce feature depth, narrow user base)
    • Increase budget to match realistic effort
    • Extend timeline (reduce team size, stretch delivery)
    • Use phased approach where Phase 1 fits budget, evaluate before continuing
  • Let SOP 4 and the client decide how to resolve the conflict
  • Flag prominently in loe-estimate.md Review Summary

Phase 1 Exceeds 12-Week Constraint

If Phase 1 estimate exceeds 12 weeks (even in most likely or pessimistic scenario):

  • STOP - This is a critical violation of the hard constraint
  • Flag prominently in every artifact Review Summary after Step 4
  • In Step 6 recommendations, provide descoping options:
    • Further reduce Phase 1 feature set
    • Narrow user base more (e.g., 50 SKUs instead of 100)
    • Reduce feature depth (e.g., remove mobile app, web-only)
    • Increase team size to compress timeline
    • Select different top-priority outcome that's achievable in 12 weeks
  • Do NOT proceed to SOP 4 until Phase 1 is under 12 weeks
  • Return to SOP 2 to descope Phase 1 if necessary

Team Context Indicates Low Experience with Key Technologies

If team lacks experience in critical technologies:

  • Apply higher novelty multipliers (1.8-2.5x instead of 1.3-1.8x)
  • Include explicit learning curve buffers (1-2 weeks per new technology)
  • Recommend technical spikes or training before Phase 1 kickoff
  • Increase "almost right" debugging overhead to 30-40% for unfamiliar domains
  • Lower confidence score (Team Experience: +5 instead of +10-15)
  • Consider recommending different technology choices that match team strengths
  • Flag in risk analysis that team expertise gap is a major risk

High Uncertainty Tasks Create Very Wide Estimate Ranges

If pessimistic estimates are >2.5x optimistic for many tasks:

  • This indicates very high project risk and uncertainty
  • Consider whether tasks need further decomposition or clarification
  • Recommend discovery phase or technical spikes to reduce uncertainty before committing
  • Use ensemble estimation methods for highest-uncertainty tasks (if time permits)
  • Apply higher confidence intervals (±50% instead of ±30%)
  • Lower overall confidence score
  • In recommendations, suggest validating high-uncertainty items before fixed-price commitment
  • Consider recommending time-and-materials or not-to-exceed pricing for high-uncertainty components

AI Tool Usage Level Is Unclear

If team's AI assistant usage is not well-defined:

  • Ask clarifying questions about proficiency level and usage patterns
  • Use conservative assumptions (assume "moderate" usage with standard debugging overhead)
  • Apply standard AI overhead: 20-30% debugging, 10-15% validation, 5-10% prompt engineering
  • Note that AI productivity varies significantly with team proficiency
  • DO NOT assume AI provides net productivity gains - research shows hidden costs negate apparent speedups
  • Document AI usage assumptions prominently
  • Recommend tracking AI-generated code quality in early iterations to calibrate estimates

Validation Reveals Estimates Are Significantly Off

If human reviewer believes estimates are too high or too low:

  • Ask clarifying questions to understand their reasoning
  • Identify which specific tasks or assumptions they disagree with
  • Adjust estimates based on their domain expertise (they know their team/domain better)
  • Update reasoning in artifacts to explain the adjustment
  • If their feedback suggests systematic bias (all estimates too high/low), adjust calibration factors
  • Regenerate affected artifacts with adjusted parameters
  • Re-present for approval
  • Remember: Human expertise should override model estimates when there's conflict
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment