Skip to content

Instantly share code, notes, and snippets.

@wilmoore
Last active February 2, 2026 17:30
Show Gist options
  • Select an option

  • Save wilmoore/c127bcc973c3d7ee19a3858a0349912b to your computer and use it in GitHub Desktop.

Select an option

Save wilmoore/c127bcc973c3d7ee19a3858a0349912b to your computer and use it in GitHub Desktop.
Business :: Ideas :: Śavvy AI :: Systems :: AI Literate :: Evidence

Business :: Ideas :: Śavvy AI :: Systems :: AI Literate :: Evidence

⪼ Made with 💜 by Polyglot.

Purpose

This document defines the canonical evidence system for AI Literate.

It specifies:

  • What constitutes admissible evidence
  • How evidence is logged
  • How evidence is classified
  • How evidence may be consumed by downstream governance systems

This document does not interpret evidence.
This document does not make decisions.
This document defines admissibility only.

Downstream systems (e.g. pricing governance, curriculum validation, authority claims) may reference this document but may not override it.


Evidence Object (Canonical Schema)

All evidence within AI Literate must be logged as a discrete object.

Evidence Item

Fields:

  • evidence_id
    Unique identifier for the evidence item

  • evidence_class
    One of the registered Evidence Classes defined below

  • source
    Human, organization, platform, or system that produced the evidence

  • timestamp
    ISO 8601 timestamp of when the evidence occurred

  • raw_artifact
    Verbatim content or direct reference
    (exact quote, link, screenshot, action, or transaction)

  • admission_status
    One of: valid | invalid | pending

  • exclusion_reason
    Required if admission_status = invalid

  • notes
    Optional. Descriptive only. No interpretation.

Global Rules

  • Evidence must be logged verbatim
  • Evidence must not be summarized
  • Evidence must not be interpreted at this layer
  • Evidence admission is binary: valid or invalid
  • Evidence may be stored globally but interpreted only locally

Evidence Class Registry

Evidence Class: Demand Signal

Definition
Unprompted expressions of intent, value, or forward demand.

Valid Forms

  • “When is the next session?”
  • Referral attempt
  • Repeat attendance without incentive

Invalid Forms

  • Survey responses
  • Facilitated testimonials
  • Prompted feedback

Evidence Class: Value Realization

Definition
Statements or actions indicating perceived value exceeding cost.

Valid Forms

  • “This was worth more than I paid”
  • Follow-on purchase without discount

Invalid Forms

  • Price comparisons prompted by facilitator
  • Retrospective survey answers

Evidence Class: Authority Signal

Definition
External validation that increases perceived credibility.

Valid Forms

  • Public endorsement
  • Named partner usage
  • Institutional reference

Invalid Forms

  • Private praise without attribution
  • Anonymous comments

Evidence Class: Retention Signal

Definition
Behavior indicating voluntary continued participation.

Valid Forms

  • Repeat attendance
  • Subscription continuation without incentive

Invalid Forms

  • Auto-renewals without engagement
  • Incentivized retention

Evidence Class: Partner Validation

Definition
Evidence originating from delivery or distribution partners.

Valid Forms

  • Unsolicited partner testimony
  • Partner-initiated referrals

Invalid Forms

  • Summarized partner reports
  • Interpreted feedback

Evidence Class: Financial Commitment

Definition
Explicit monetary action indicating commitment.

Valid Forms

  • Full-price purchase
  • Upgrade without discount

Invalid Forms

  • Discounted purchases used for experimentation
  • One-off goodwill payments

Consumption Rules

  • Evidence classes are defined centrally
  • Governance modules may reference evidence classes
  • Governance modules may define thresholds
  • Governance modules may not redefine admissibility

Change Control

  • This document is treated as system infrastructure
  • Changes require explicit versioning
  • Ad hoc edits are prohibited

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