LEADERSHIP JUDGMENT CHALLENGE 006
THE CONTROL
THAT PASSED
A procurement review ran for eighteen months. Every required review was completed, recorded and signed, with zero missed steps. Now procurement wants to use that 100% figure as the evidence for raising the threshold. “It’s still 100% compliant,” says the report. “Compliant with what?” says the auditor.
What would you do?
THE SITUATION
Every review was completed. What does that actually prove?
A procurement review process has run for eighteen months. Every required review was completed, recorded and signed. Zero missed steps.
An audit looks closer. Most of the records show no evidence that anyone examined what mattered. The reviews happened. The question is whether they did anything.
Procurement now wants to raise the threshold, using the 100% figure as the evidence. That is why what the record actually proves matters now.
The report says: “It’s still 100% compliant.” The auditor replies: “Compliant with what?”
MAKE THE CALL
What do you do?
Choose before you continue.
WATCH THE CHALLENGE
Coming January 28.
The video for this challenge publishes January 28. Check back then — or work through the situation and questions below in the meantime.
THE JUDGMENT PROBLEM
This isn’t about a mistake. It’s about a record doing its job.
Nobody skipped a step. The control recorded exactly what it was designed to record: that each review happened.
The question isn’t “did the reviewers do something wrong?” It’s: what does a 100% completion figure actually prove — that a review happened, or that a review worked?
A completed control tells you a step happened. It doesn’t tell you whether the step did what it was built to do.
The control didn’t fail. It’s the difference between a completion record and an effectiveness record, becoming visible because someone now wants to rely on it.
THE EVIDENCE PROBLEM
Completed is not the same as effective.
More compliance doesn’t answer the question. It just restates it.
AI can perform perfectly. Judgment determines whether the control around it deserves your confidence — and whether that confidence has ever been tested.
THE JUDGMENT DIFFERENCE
A completion record is not an effectiveness record.
That a step happened.
Did anyone examine what mattered?
The control passed. Whether it proved anything is a different question.
BETTER QUESTIONS
Before you expand the threshold, ask:
Has this gap been there all along, or is this genuinely the first time anyone looked?
Would you cite this record as evidence of safety, knowing what it actually measures?
What is the 100% figure actually evidence of: that review happened, or that review worked?
This isn’t about which option is correct. It’s about knowing what your evidence can actually carry.
SO, WHAT WOULD I DO?
Each option is defensible — depending on how much uncertainty you can accept while the evidence catches up.
A uses the AI’s own track record to keep moving — but extends authority on a record that proves completion, not effectiveness. B protects the standard of evidence — but holds up an expansion the track record might support. C builds a real test before deciding anything — but delays a decision procurement is already pushing for. D limits the exposure to standard transactions while the test is piloted — but expands authority before the test has said anything.
This isn’t about which option is correct. It’s about knowing what your evidence can actually carry.
THE LEADERSHIP JUDGMENT TAKEAWAY
A completed control tells you a step happened.
It doesn’t tell you whether it did what it was built to do.
Before extending authority, ask for evidence that what mattered was examined.
A completed control tells you a step happened.
It doesn’t tell you whether it did what it was built to do.
BRING THE CHALLENGE TO YOUR TEAM
Make judgment visible.
Judgment Challenges can be explored with your team through a private 90-minute Judgment Lab, turning the scenario into a practical conversation about controls, evidence and what a completion record actually proves.