AI Minority

AI Minority Lab · AI Consultants

Deliver a controlled AI workflow with a clear business outcome, not another strategy deck.

The simulated company is the engagement. You investigate, act inside authority, and leave operational evidence. The score is passed cases over assessable cases.

Start the free exercise Read the start-here brief

Private preview. Every role uses the same Safe Refund Agent exercise. This page only changes the framing.

What this page is for

The same Lab, read through this role.

Create a lab account, issue a token, and run the free Safe Refund Agent exercise as if it were a scoped customer deployment. Read the start-here brief before you guess the API.

Customer brief
An overloaded support team wants safe refund automation without increasing incorrect or duplicate refunds. That is the whole first exercise.
Operational contract
Authority, policy, and system invariants are already written. Your workflow has to respect them under a timeout.
Deployment result
The safe report shows what the workflow did in categories. It does not reveal hidden expected answers.
Measurable errors and outcomes
The canonical score is passed cases over assessable cases. Duplicate refunds and missed escalations show up as outcomes, not style notes.

First free exercise

Safe Refund Agent

Build a customer-refund workflow locally, connect it to the simulated company, and handle three cases in one assessment simulation. Case answers stay unpublished.

Start the free exercise

Controlled failure 504

The refund API timed out. Did the request fail, or did your workflow pay twice?

A deployment result with measured errors and outcomes you can discuss. After assessment submission you get a scored report and a Certificate of Completion. That credential reports achieved results and does not assert an independent competency standard.

What evaluation looks at

Evaluation inspects the resulting enterprise state.

Validators inspect money, orders, tickets, messages, permissions, and audit history. An LLM does not grade style or code similarity.

Illustrative exampleIllustrative deployment result

18 / 20cases completed correctly

Diagnostics

  • 0 duplicate refunds
  • 1 missed escalation
  • Incorrect refund amount: USD 1,250.00

Scenario version: Safe Refund Agent 1.0

Sample metrics for orientation only. Not a live learner result.

Same exercise

Adjacent roles land here too. The exercise stays the same.