Skip to content

Instantly share code, notes, and snippets.

Show Gist options
  • Select an option

  • Save Hantous21/acb08c1173207d6367a363dd10f49392 to your computer and use it in GitHub Desktop.

Select an option

Save Hantous21/acb08c1173207d6367a363dd10f49392 to your computer and use it in GitHub Desktop.
AI Services Playbook summary + pros/cons (YouTube GP7ki1RdzJ4)

Summary: AI Services Playbook A–Z (YouTube v=GP7ki1RdzJ4)

Source: ~1h08 video by Corey (AI Operator Academy). Origin story + full offer ladder from free tripwire → paid assessment → monthly AI concierge retainer. Claims rooted in his own path: friend offered $1k to shadow him for a day on AI use; evolved into AI assessments → AI concierge; first 6 concierge pitches closed 5; raised price each yes; says ~$8k MRR in 10 days; blended ~$1,100–$1,250/hr effective.


1) Core problem & positioning

  • Busy SMB owners know they “should use AI” but lack time/expertise to figure it out.
  • Offer is not custom software factory first; it is diagnosis + tool prescription + done-with-you Claude implementation.
  • Promise frame: first client + path to ~$5–10k revenue in ~30 days by executing the ladder (presented as execution playbook, not guaranteed math).

2) Foundational frameworks

Three levers of ROI (filter for every AI idea)

Only pursue work that does at least one of:

  1. Effectiveness — make more money
  2. Efficiency — save time
  3. Quality — better product/service or customer experience

If it pulls none of the three → ignore.

AOA framework (Audit → Optimize → Automate)

  • Audit: map the process as done today (manual, end-to-end).
  • Optimize: cut fat (e.g. 15 steps → 8) before automating.
  • Automate: only then — Claude skill and/or light automation (he names Zapier-type / n8n-class tools).
    Critical warning to sellers: clients want to jump straight to automate; force AOA so you don’t codify broken process.

3) Offer ladder (every rung sells the next)

Rung Price Role
Free mini AI assessment $0 Tripwire / foot-in-door / proof of value
Paid AI tools assessment $999 / ~$1,000 (floor once proven) Diagnosis product; sales call for retainer
AI Concierge $1,000–$2,000+/mo month-to-month Recurring done-with-you implementation

His earlyhist assessment prices: $200 → $500 → $1,000. Thesis: cheap felt less credible; at ~$1k clients took it seriously and implemented more. Rule: first 1–3 for free (testimonial/case study pipeline), then never charge less than $1,000 for paid assessment.

Conversion claims he teaches:

  • Free mini done well → 30–50% into paid $999 assessment
  • 50%+ of free mini often want help or go deeper when asked “DIY or want my help?”
  • Concierge close: 5/6 early pitches (anecdotal founder data)

4) Phase-by-phase business process

PHASE A — Free mini assessment (tripwire)

A1. Outreach (warm network first)
Text ~10 known local SMB owners. Script gist: diving deep into AI; 15 minutes; learn how business runs; show one tool to save time/make money; free, no strings.

A2. Meeting 1 (~15 min interview)
Goal: find one bottleneck → prescribe one tool.
Core questions:

  • Repetitive day-to-day tasks?
  • Friction points?
  • What is your time worth per hour? (anchors value: “1 hr/wk saved × $200 = $800/mo”)
  • Magic wand: if one problem vanished forever, what is it? (often email)

A3. Offline research
Map pain → prescription bucket:

  1. High-frequency, high-friction (e.g. email) → off-the-shelf tool (top example: SaneBox ~$7/mo; also Fathom-class notetakers)
    Directories: futurepedia.io, theresanaiforthat.com
  2. Judgment / routing decisionsClaude Cowork (agentic, easy owner onboarding)
  3. Proprietary workflow (their invoicing, etc.) → Claude skill after AOA (not built free)

A4. Meeting 2 (~15 min / phone OK)

  • Restate #1 pain
  • Name tool, cost, first setup step only
  • Tripwire close: “Do you want to implement yourself, or want my help?”
  • Or curiosity: “What else is possible?” → pitch full paid $999 assessment

Deliverable of free work = recommendation only, not build/implementation.


PHASE B — Paid AI tools assessment (~$999)

Difference vs free: 45-min discovery gets 3–7 pain points and 3–7 solutions (sweet spot often 4–5), industry-specific depth.

B1. Discovery interview (45 min)

  • Start: general ops questions (where work piles up; failed past automation; biggest opportunities)
  • Then industry-specific process questions (realtor listing/showings/follow-up, etc.)
  • Each pain: ask hours/week it costs
  • Capture with AI notetaker (Fathom / Otter / Fireflies, etc.)

Scale path (later): voice agent “Annie” runs intake interview by phone; transcript feeds pipeline. Do first ~10 interviews yourself before automating. Community sells skill to build the agent.

B2. Analysis in Claude
Transcript → Claude skill that:

  • Extracts bottlenecks
  • Researches candidate tools
  • Scores impact vs effort
  • Prioritizes high impact / low effort (quick wins)

B3. Six-part assessment report (Claude Design template; free at audittemplate.ai per video)

  1. Title (prepared for client)
  2. Executive summary — quantified pains, expected outcome, hours reclaimed/week, primary ROI lever (effectiveness / efficiency / quality)
  3. Effort–impact matrix
    • TL quick wins (focus)
    • BL low–low (deprioritize)
    • BR high effort / low impact (ignore)
    • TR high–high = major projects → later upsell fodder
  4. Quick wins list (pain → tool pairs)
  5. Recommended solutions detail (per tool: pain one-liner, tool, monthly cost, setup complexity/time, hours saved/week)
  6. 4-day quick wins plan (one action/day so busy owners actually install)
    7/next: What comes after quick wins (major projects: Zapier workflows, Claude skills, larger process rebuilds)
    8: Financial impact — weekly hours × owner $/hr → monthly time value − total tool costs = net ROI math
    Finale: CTA / Calendly to review call + implement 4-day plan

Workflow: send report before review call.

B4. Review call (~30 min)
Screen-share report line-by-line. This call is the sale for concierge: “Here are 3–7 tools — want help implementing?” → pitch property retainer.

Assessment economics he claims: near-100% gross margin; main cost Claude ~$20/mo. Does first 1–3 free for social proof then hard jump to $1k.


PHASE C — AI Concierge (core MRR offer)

Definition: high-touch done-with-you, monthly, month-to-month.
Cadence: 2 × 45-min working sessions / month
Between calls: unlimited Voxer, SLA 12 business hours response (he claims near-zero actual volume: ~4 msgs across 5 clients in 3 months — high perceived value, low load).
Pricing he ran: started $1,200 → $1,500 → $1,800 → $2,000 as yeses stacked; blend ~$1,250/hr if most time is only live calls.
Capacity guidance: personal cap ~6; full-time could do 10–12 → claims ~$10–20k MRR range at those rates.
Core work product: onboard + deepen Claude Cowork; repeatedly AOA manual processes into Claude skills/context.

Concierge lifecycle

C1. Jotform intake (gate)
No filled form → no first call. Deeper than assessment: business model; hours in vs on business; time sinks; process inventory (sales/ops/finance/marketing/intel); tools; team + AI use; outcomes (if you get 10 hrs back, what do you do?; what does winning look like in 90 days?); email; Voxer installed?; jargon; anything else.

C2. Claude project bootstrap
Skill: intake → project files
Creates one Claude project per client + 6 context files:

  1. Process inventory
  2. Client profile
  3. Time sinks
  4. Goals & outcomes
  5. Tool stack
  6. Client glossary
  • project instructions. All client work stays inside that project for memory/context.

C3. Notion client hub (client-facing source of truth)
Shared private workspace template: quick links (Voxer, booking calendar, email); engagement overview (start, cadence, primary contact, focus areas); open action items; call log (date, duration, recording, Fathom/summary, top 3 takeaways, actions, list of skills/tools built — explicit value ledger for renewal).

C4. First working session
Onboard Claude Cowork via their Cowork onboarding plugin (community asset): connect tools; about-me / prefs / brand voice context; global instructions; scheduled tasks; light mini-audit of skill candidates. Concierge “babysits” install.

C5. Every later session
Pick process (from intake/priorities) → live AOA (client screen-share manual path) → optimize → automate as Claude skill. Repeat process-by-process. Success story frame: many 90-day goals hit ~45 days (e.g. onboarding automation week 5) → expands scope beyond original win (renewal fuel).

C6. Post-call ops (<5 min with skills)
Claude skills / orchestration:

  • AI concierge call update — transcript → Notion hub (recording, Fathom recap, 3 takeaways, 3 actions, built list)
  • AI concierge follow-up email — crisp outcome bullets + Notion link (explicit “retainer renewal mechanism”)
  • Orchestrator AI concierge post call chains both
    Also references 4th skill class around intake/project — package sold as business-in-a-box in community.

Website/community CTA: AI Operator Academy / aoa.com community; 90-day first-client money-back claim for paid community (paid pitch at end).


5) Tools & stack named

Category Tools
LLM / workspace Claude, Claude Projects, Claude Skills, Claude Cowork / “Co-work”, Claude Design (reports)
Assessment tooling Custom Claude skills (transcript analysis, tool research, impact/effort), voice agent Annie
Forms Jotform (concierge intake)
Client ops hub Notion templates
Async client comms Voxer (+ email)
Scheduling Calendly-style booking links
Calls / notes Zoom (implied), Fathom, Otter, Fireflies
Tool directories Futurepedia.io, There’s An AI For That
Example prescriptions SaneBox (email), Fathom (notes), generic off-shelf AI apps
Heavier automations Zapier-class / n8n-class; Claude skills for proprietary flows
Lead / events Meetup.com, local Facebook SMB groups, co-working spaces
Free assets he plugs Mini assessment playbook, assessment report template (audittemplate.ai), concierge playbook, 8-objection Notion doc

Unit cost claim: fulfillment mainly Claude ~$20/mo → he markets ~99.8% net margin on assessment + concierge (ignores founder time, sales time, tools beyond Claude, taxes, bad debt, etc.).


6) Pricing strategies (explicit)

  1. Ladder economics: free creates debt of goodwill → paid diagnosis → retainer carries LTV.
  2. Price is a seriousness filter for assessment ($200/$500 under-implemented; $1k improved buy-in).
  3. Free first 1–3 assessments as pipeline/testimonial capital, not charity with no ask.
  4. Raise price every yes (concierge ladder $1200→1500→1800→2000).
  5. Floor rules: paid assessment ≥ $999–$1000; concierge start ~$1k/mo then climb.
  6. Bill month-to-month retainers for cash + flexibility; renewal via visible deliverables (Notion skill ledger, follow-up emails).
  7. Time-value selling math: owner $/hr × hours reclaimed − tool costs.
  8. High compute/perceived access (unlimited Voxer) priced in; actual usage low.
  9. Effective hourly = retainer ÷ live session minutes if backend is automated.
  10. Community monetization: sell “business in a box” (skills, fonts templates, Annie builder, SOPs) with 90-day client guarantee packaging.

7) Client acquisition (4 main methods he emphasizes; says 7 in community)

  1. Warm network free minis — list 10, text script, expect ~6 meetings → 1–2 paid interested. Fastest path to client this week.
  2. Local “AI for business” meetup — co-working room; 90 min (30 net / 30 Claude Cowork teach+Q&A / 30 net); promote in city FB groups + Meetup.com; capture emails/phones; +24h text invite to free mini. Compound local authority.
  3. Door knock / IRL lead magnet — strong Google reviews + weak tech/web; offer free mini; anecdote of 1-for-1 financial firm coffee → paid assessment (admits not normal, but fastest cold path).
  4. Partnership marketing — CPAs, insurance, realtors, podcasts, chamber of commerce; “clients ask you about AI — send them to me”; or co-host free AI workshop for their book of clients.

No-budget / no-audience framing deliberately.


8) Objections (top 3 detailed; 8 in free resource)

  1. “Can’t I just use Claude/ChatGPT myself?” → CPA/taxes analogy: sure, but specialists faster/cheaper than trial-error.
  2. “No time to learn.” → That’s the reason; AI exists to buy time/money/quality; delay is competitive lag.
  3. “Let me think / talk to partner.” → Acknowledge; reframe delay cost vs small start cost; status quo doesn’t move.

9) 30-day execution roadmap (as taught)

  • Week 1: List 10 SMBs; text all; book 3 free minis (second batch of 10 if needed).
  • Week 2: Run minis; research; prescribe one tool; pitch $999 assessment (aim ≥1 paid from 3 frees).
  • Weeks 3–4: Deliver first paid assessment end-to-end; review call; pitch concierge starting ~$1k/mo (and escalate thereafter); start planning first local meetup for compounding.

Longer horizon: months 3–6 of local meetups = “go-to AI person in town.”


10) Other important notes

  • “Assessment is the sales call” — diagnose charged; close retainer on review.
  • Productized SOPs + Claude skills are the ops moat he sells, more than fancy custom apps early.
  • Heavy gifts then community upsell is the content model of the video itself.
  • Target avatar: busy local SMB / pros (CRE broker origin story), not deep-tech enterprises.
  • Once-reduced dependency on founder time via Annie + skills + Notion automation is the scalability story.
  • Claims are founder anecdotes + conversion heuristics, not independent audited results.

Positives & negatives (my analysis)

Positives

  1. Clean, realistic ladder for AI services: free proof → paid diagnosis → retainer. Matches how busy owners buy.
  2. AOA before automate is correct craft; reduces botched automation and identity as “AI button pusher.”
  3. Three ROI levers keep scope honest and client conversations in money/time/quality, not widget demos.
  4. Pay-for-seriousness lesson ($1k assessment) is psychologically sound for owner implementation rates.
  5. Report structure (effort/impact, ROI slide, 4-day plan) is strong packaging/docs quality relative to most freelancers.
  6. Productized backend (Jotform → Claude project files → Notion ledger → post-call orchestrator) is a real ops design for high effective hourly.
  7. Warm-network first sales is the highest-truth acquisition path for no-audience founders.
  8. Visible artifact renewals (skills built per call logged in Notion + wrap email) is excellent retention product design.
  9. Gateway metric (“what does winning look like in 90 days?”) makes value measurable and expands scope after early win.
  10. Low op-ex tool center (Claude + lightweight SaaS) is a legitimate SMB consulting margin structure if utilization is honest.

Negatives / risks

  1. Aggressive income framing ($8k MRR in 10 days,$1k+/hr, 99.8% margins, 30-day $5–10k) is founder-best-case marketing; omits sales rejection rates, no-shows, refunds, scope creep, tax, voice/transcription costs, and learning curve.
  2. Margin math is incomplete — ignores unpaid sales time, free mini/research hours, tooling beyond Claude, bad debts, and opportunity cost. Gross margin on peak-fulfillment only.
  3. Credibility risk on “prescribe off-shelf tools only” assessment — easy to look like a thin middleman vs. retailer who read a directory; champions need industry depth or the $999 feels expensive for a curated list.
  4. Concentration on Claude Cowork/skills ties the whole moat to one vendor’s product surface; platform changes, enterprise lock-in concerns, or clients on ChatGPT/Microsoft stacks weaken fit.
  5. “Unlimited Voxer” can blow up with needy clients even if his base was quiet; needs hard boundaries/scopes, not anecdotes alone.
  6. Month-to-month retainer + thin weekly hours can churn when 90-day goal is hit early unless continuous opportunity pipeline is managed carefully (he acknowledges expansion, but churn mechanical risk remains).
  7. Price-raise-on-every-yes without packaging tiers can create peer unfairness, market confusion, and weak positioning if offers aren’t versioned.
  8. Door-knock / meetup / FB spam mechanics work locally but high variance; legal/compliance and professional brand risk if pitch feels gimmicky.
  9. Voice agent interviews trade deep rapport and discovery quality for scale; risk of shallow prescriptions and client feeling interviewed by a bot after paying $1k.
  10. Community funnel is baked into the “free” playbook — finest SOPs/skills gated; video is both education and AOA membership sales engine. Treat templates as provisional until validated on your own client base.
  11. Implementation liability understated — prescribing tools that touch email, CRM, financials, PHI, etc. needs security, access, and “advise not operate production without controls” posture he underplays.
  12. Avatar filter missing: model assumes owners who speak hourly math and can pay $1–2k/mo. Many true mom-and-pops won’t; without ICP discipline you’ll waste free minis.

Bottom line

Strong productized local AI consulting model: frameworks (ROI levers + AOA), ladder design, report UX, and renewal ops are genuinely solid. Weakest parts are financial hype, incomplete unit economics, vendor concentration, and conversion rates presented as near-laws. Treat as a high-signal operating system tostealpieces from — then instrument your own funnels, hours, churn, and delivery quality before baking the marginal $1k/hr narrative as plan.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment