About this gut check
What this gut check measures
The gut check gives you two scores out of 100. Gut trust is the number you pick on the first screen: how much you trust your AI agent. Provable trust comes from your answers to the questions that follow, and reflects how much of that trust you could show to someone else today. The gap between the two is the point of the exercise.
Provable trust rises when trust rests on evidence: tests you could show an auditor, tests you can re-run when the model changes, monitoring against defined thresholds, a named owner, and an inventory of the AI systems you run. It falls when trust rests on a good demo, a vendor's claim, a reviewer who mostly clicks approve, or the fact that nobody has complained yet.
A large gap does not mean your agent is unsafe. It means that if someone asked you to prove it is safe, you could not do it today. Closing the gap is about evidence: tests, approvals and monitoring records, kept where reviewers can find them.