Answerable for what you didn’t make.

A team I backed walked into a room with leadership and got taken apart. They’d been fully in sync with the product owner, felt they had market fit, and designed something that made the best of a complex, anti-customer policy.

Leadership asked why nobody had pushed back on the policy itself.

I’d backed that team through every stakeholder session. I assumed someone upstream had checked whether the policy was fixed or just unquestioned. I hadn’t checked that assumption. We’d put the work into solving around it.

In the room, I stepped in front of the team and said it was my call. We’d made assumptions that were being proved wrong, and my recommendation was to go back and challenge the policy.

The lead client thanked me for the honesty. He said he would have appreciated us pushing back from the customer’s perspective, advocating for the audience that wasn’t in the room. He shelved the project for the moment and went to fight to change the policy.

I’d been close to the work the whole time. Being aligned with the people in the project had let me assume we were doing right by the people it was for.

The part I still have to check.

I think about that room when I work with AI. I can hand it an assumption and get something useful back before I’ve spent much time questioning the assumption myself. More passes can make the work better while leaving that question untouched.

A model can help me challenge a premise. I still have to decide whether we’ve examined it properly, including whose perspective is missing. “The model didn’t flag it” would have been no better an answer in that room than my assumption that someone upstream had checked.

The temptation is familiar. I want to get into the work, find the solution, make it sharper. When I can get a first pass back quickly, I have more room to question what I’ve asked for. I can also spend all of that time refining the answer.

In that client meeting, I owned the call after the work had been challenged. The client was asking for something earlier: bring the customer’s perspective into the decisions, while there was still a chance to question what we were making.