ARCHAI PROOF OF JUDGMENT — LIVE CHALLENGES
You approved the AI agent.
What exactly did you approve?
The policy exists. The assessment is complete. The approval is recorded.
Now prove that the evidence covers the version, permissions and actions being authorized.
THE SCENARIO
The File Looks Complete. Is the Decision Justified?
A fictional organization wants to deploy an AI customer-service agent that can access account information and request financial adjustments.
Participants receive a policy, risk assessment, test summary, approval record and proposed operating scope.
Their task is to determine whether the requested authorization is supported.
The main event's critical discoveries remain undisclosed on this public page.
Would You Sign This Approval?
This exercise runs locally in your browser, requires no registration and does not transmit or retain your answers.
Version 1.2 tested in read-only mode.
Version 1.3 requests permission to modify records.
"Approved for pilot." No explicit scope is recorded.
Would you sign this approval?
This illustrative exercise does not provide a professional certification or a governance maturity score.
THE LIVE EXPERIENCE
Six Phases, One Authorization
- 01Read the authorization request.
- 02Inspect the evidence.
- 03Record your initial decision.
- 04Examine new information.
- 05Resolve objections and define conditions.
- 06Present the decision dossier.
Revising a decision in response to evidence can demonstrate strong judgment.
WHAT PARTICIPANTS WILL BUILD
You Will Work On
- Requirement-to-action worksheet
- Evidence register
- Version and permission comparison
- Objection and resolution log
- Approval conditions
- Decision dossier
- Follow-up action register
These are proposed deliverables you will work on — they are not guaranteed to be complete regardless of participation.
THE EVIDENCE LAYER
A document says a control exists.
Evidence helps establish whether it operated.
REQUIREMENT
What must be satisfied?
ACTION
What work addresses it?
OWNER
Who is responsible?
EXECUTION
What actually happened, when and on which version?
EVIDENCE
What record supports the claim?
CHALLENGE
Who objected, and how was the objection handled?
DECISION
Who authorized what, why and under which conditions?
REVIEW
What change requires a new decision?
A generated report does not prove that a control was executed. Missing evidence must remain visible.
This is the working model participants will practice. It does not imply that an automated evidence repository has already been deployed, and it does not fabricate tool names, API endpoints or native Studio integrations.
FACILITATOR
Leonardo Ramírez
Enterprise Architect, AI Governance Practitioner and Founder of ARCHAI WORLD.
Leonardo connects enterprise architecture, AI governance and mission-based execution to help professionals examine decisions with context and evidence.
Already part of YOU ARE THE ARCHITECT?
Check your cohort's included experiences before purchasing separately.
Inclusion of this challenge does not automatically apply to every cohort.
Go to YOU ARE THE ARCHITECTBring a decision your organization needs to defend.
Discuss a private facilitated exercise using a synthetic or authorized case, with a separately agreed scope.
DISCUSS A PRIVATE CHALLENGEREGISTER YOUR INTEREST
Notify Me When Registration Opens
Please describe the issue at a high level. Do not submit confidential company information.
Frequently Asked Questions
Architects, CIOs, AI governance leaders and professionals responsible for consequential decisions who want to examine whether an AI approval is supported by evidence.