Free decision tool

Pressure-test the idea before you buy the AI.

Turn an AI idea into a decision brief your team can challenge. The stress test surfaces missing evidence, unclear ownership, failure exposure, and the smallest responsible test.

Start the stress test

How it works

A useful decision is more than a readiness score.

The tool checks whether the problem, data, decision exposure, ownership, controls, and evidence are clear enough to support a next step.

It does not approve an AI system or replace security, privacy, legal, accessibility, or domain review.

Build your brief

Describe the work, not the technology.

Required fields establish the use case. Leave optional fields blank when the answer is unresolved—the brief will treat that as a gap.

01Define the work

Use case name

Use a working label, not a vendor or product name.

Who uses it or experiences the result?

Name the employees, customers, partners, or communities affected.

What task should become easier or better?

Describe the work in plain language.

What useful outcome should change?

Describe an observable result, not “use AI” or “increase efficiency.”

02Inspect the dependencies

What information would the system need?Optional

Describe categories, sources, quality, and access. Do not paste real customer, employee, or confidential data.

Could the use case involve personal, confidential, regulated, or safety-sensitive information?

No known sensitive informationBased on the intended scope and available data.Not sure yetThe data or downstream use is still unclear.YesThe use case may involve protected or high-impact information.

What decision or action could the output influence?Optional

Include decisions made by people, workflows, or automated systems.

What happens if the result is wrong?

Low consequenceAn error is easy to notice and correct.Moderate consequenceAn error creates rework, delay, confusion, or a poor experience.High consequenceAn error could materially affect a person, commitment, right, or business outcome.

How reversible is the AI-influenced action?

Easy to reverseNo lasting effect if corrected quickly.Partly reversibleCorrection is possible, but some effect or cost remains.Difficult to reverseA wrong action may create lasting harm or obligation.

03Define ownership and evidence

Who owns the outcome when the system is wrong?Optional

Name a role with authority to intervene—not “the business” or “the AI team.”

Where does human judgment enter the workflow?Optional

Describe who reviews what, when they can stop it, and what evidence they see.

What evidence would show the use case is useful and safe enough?Optional

Include quality, customer or employee impact, error patterns, and operating cost.

What is the smallest useful test?Optional

Prefer a limited sample, reversible setting, named owner, and clear stop condition.

What it considers

Six questions behind the brief.

01
Problem Is the work and useful outcome clear?
02
People Who experiences the decision and its errors?
03
Data Is the necessary information known, appropriate, and available?
04
Exposure What happens when the output is wrong?
05
Ownership Who can review, intervene, and answer for the result?
06
Evidence What would justify continuing, changing, or stopping?

Pressure-test the idea before you buy the AI.

A useful decision is more than a readiness score.

Describe the work, not the technology.

AI use case

The work being considered

Dependencies to resolve

Ownership questions

What the test must challenge

A bounded way to learn

Pause or stop the test when

Six questions behind the brief.