Twelve questions. Two minutes.

Could three different people on your team give the same answer to "what are our top three priorities right now"?

In the last quarter, did you formally kill or say no to a planned initiative?

Of what you committed to ship last quarter, did more than two-thirds actually ship?

Could a new engineer build the right thing from a typical ticket without a follow-up conversation?

Does someone specific own the quality of specs and tickets — a named human, not "the process" and not the AI?

When work gets built wrong, do you usually know within days (not weeks)?

Is less than a quarter of your team's completed work unplanned (injected mid-sprint or mid-cycle)?

Of the process you adopted in the last year, is most of it still actually followed today?

Can your engineers name what was deliberately removed from the process in the last year — a meeting killed, a field deleted, a gate dropped?

Does AI-generated work product (tickets, specs, code) pass through a human gate that genuinely rewrites, not rubber-stamps?

Could you say, with evidence, whether AI tooling has made your team faster or just busier?

When something ships, do the people who talk to customers reliably find out within the week, in a form they can use?