Skip to resource
MHBMMichael Hanna-Butros MeyeringComplex systems · human outcomes
PrivacyAI disclosureConnect
Menu
Navigation fallbackExplore MichaelHBMAvailable even when interactive navigation is still loading.
Connect
NavigationExplore MichaelHBM
Connect

30 checks · six receipts · one accountable disposition

Six receipts before scale.

A printable AI change management review for one real workflow. Check the operating evidence—not whether a slide, policy, training, or launch meeting exists.

Practice
Public-sector technology · privacy · responsible AI
Updated
August 4, 2026
Use
Source-backed · printable · adaptable
Reusable artifact

Use it in a working session, then attach the evidence behind every answer.

Use in a working review

Choose one workflow. Bring the evidence.

Complete the review with the people who own the service, technology, affected work, controls, and support. A checked box means the linked evidence supports the statement—not that someone remembers discussing it.

Exit condition
Record the owner, approved boundary, open gap, receipt link, disposition, decision date, and next trigger. If the six receipts cannot be reconciled, keep the affected path supervised or paused.
Checklist status

This tool adapts Michael's published six-part operating framework. It is intentionally stricter than a launch-readiness list: any material safety, rights, privacy, security, accessibility, continuity, ownership, or rollback failure remains a hold even if most boxes are checked. Replace its starting guardrails with your applicable requirements and approved risk tolerance.

Interactive + printable worksheet

Build the decision record as you review.

Your entries stay in this browser page and are not submitted or stored by the site. Print or save the completed page before leaving if you want to keep it.

01

Detect the operating change

Name the change and the real decision it creates.

Required receipt

Dated signal + affected workflow + service baseline + decision owner and deadline.

Hold when

The problem, population, baseline, or accountable owner is unnamed.

02

Name the owner and boundary

Make authority, identity, data, and limits inspectable.

Required receipt

Named owner + explicit allowed, denied, and approval-required boundary + required approvals.

Hold when

Effective permissions, data movement, high-impact decision rights, or final accountability cannot be inspected.

03

Redesign and test the workflow

Prove the real task, handoffs, exceptions, and human judgment.

Required receipt

Versioned workflow + test set + baseline comparison + exceptions and corrective actions.

Hold when

Critical accessibility defects, unmitigated high-impact errors, or materially worse task outcomes remain.

04

Prepare people to operate it

Give each role the reason, practice, support, and voice it needs.

Required receipt

Role guidance + observed practice + support ownership + consultation and readiness gaps.

Hold when

A role with material responsibility has no guidance, successful practice, or reachable support path.

05

Prove adoption and control

Review behavior, value, rights, and operating health together.

Required receipt

Joined adoption + outcome + rights/control + operating-health review with disposition.

Hold when

Usage is the only success measure, critical failures are averaged away, or no owner can decide the next state.

06

Keep rollback real

Prove the path can stop, correct, recover, and communicate.

Required receipt

Witnessed rollback + elapsed time + correction scope + retest + final disposition.

Hold when

A critical rollback step is untested or fails, affected work cannot be found, or continuity depends on the failed system.

Final operating decision

Issue the disposition.

Select one state for the named boundary. Do not use “approved” without duration, population, data, actions, residual risk, decision owner, evidence, and a trigger to review again.

Reusable artifact

Move the checklist into your operating system.

Copy the plain-text version into a ticket, decision memo, document, or governance record. Preserve the language that a checked item requires evidence.

AI CHANGE REVIEW — SIX RECEIPTS BEFORE SCALE

Workflow / decision ID:
Accountable owner:
Affected people / service:
Review date:
Next trigger:

01 DETECT THE OPERATING CHANGE
[ ] Change and date named
[ ] Workflow and baseline documented
[ ] Affected people identified
[ ] Owner and decision deadline named
[ ] Non-AI alternative considered
Receipt link:
Open gap / owner / due date:

02 NAME THE OWNER AND BOUNDARY
[ ] Allowed, denied, and approval-required actions explicit
[ ] Identity, credential, connection, and disposition reconcilable
[ ] Data purpose, movement, retention, correction, and deletion recorded
[ ] Required review owners engaged
[ ] Vendor, change-control, continuity, and exit terms bounded
Receipt link:
Open gap / owner / due date:

03 REDESIGN AND TEST THE WORKFLOW
[ ] Before-and-after workflow complete
[ ] Normal, edge, accessibility, misuse, and adverse cases tested
[ ] Outcome compared with baseline
[ ] Human review, correction, appeal, and incident path tested
[ ] Limitations visible at point of use
Receipt link:
Open gap / owner / due date:

04 PREPARE PEOPLE TO OPERATE IT
[ ] Role guidance and authority clear
[ ] Successful task and exception practice observed
[ ] Support and oversight responsibilities assigned
[ ] Required consultation / engagement complete
[ ] Support, complaint, correction, and feedback route ready
Receipt link:
Open gap / owner / due date:

05 PROVE ADOPTION AND CONTROL
[ ] Eligible, enabled, activated, and retained use separated
[ ] Task / service outcome compared with baseline
[ ] Overrides, errors, complaints, appeals, incidents, and impacts reviewed
[ ] Cost, support, drift, evidence, and correction time visible
[ ] Owner issued a dated disposition
Receipt link:
Open gap / owner / due date:

06 KEEP ROLLBACK REAL
[ ] Access, credentials, tools, and configuration can be revoked
[ ] Affected work can be found, preserved, corrected, and reprocessed
[ ] Export, retention, deletion, and vendor exit tested
[ ] Prior safe process restored within target
[ ] Notification and support path ready
Receipt link:
Open gap / owner / due date:

FINAL DISPOSITION
[ ] Scale  [ ] Proceed narrowly  [ ] Revise  [ ] Pause / contain  [ ] Retire
Critical hold open? [ ] No [ ] Yes:
Boundary / residual risk:
Decision owner / date:
Evidence index:
Next review trigger:

Method, limits + sources

A working aid—not a substitute for accountable review.

Michael’s practitioner synthesis connects operating change, public-sector delivery, and the six receipts before scale. Every organization remains responsible for applying its own authority, expertise, evidence, and risk tolerance.

Limitations

  • This is practitioner guidance, not legal, audit, labor-relations, procurement, records, privacy, security, civil-rights, or accessibility advice.
  • The starting thresholds are operating guardrails, not universal benchmarks. Replace them with the applicable law, policy, risk classification, service baseline, collective-bargaining obligation, and tolerance approved by your organization.
  • A completed template is not evidence by itself. Attach source records, test results, approvals, observed outcomes, and a final disposition.
  • Do not average away a critical failure. A material safety, rights, privacy, security, accessibility, or mission-continuity gap remains a stop condition even when the overall score looks strong.

Primary and public sources

  1. Artificial Intelligence Risk Management Framework (AI RMF 1.0)National Institute of Standards and TechnologyVoluntary, rights-preserving framework for governing, mapping, measuring, and managing AI risk.
  2. NIST AI RMF PlaybookNational Institute of Standards and TechnologySuggested actions and documentation practices; NIST explicitly describes it as neither a universal checklist nor an ordered set of steps.
  3. Artificial Intelligence: An Accountability Framework for Federal Agencies and Other EntitiesU.S. Government Accountability OfficeAccountability practices organized around governance, data, performance, and monitoring.
  4. Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence ProfileNational Institute of Standards and TechnologyCompanion profile for risks that are distinctive to or intensified by generative AI.

Update history

Versioned in public.

  1. Initial publication: 30 operating checks, six required receipts, six hold conditions, interactive and printable review fields, plain-text template, primary sources, and explicit limitations.