We wrote down every question someone asked us about AI while they were trying to use it: in training sessions, discovery calls and 1:1 set-ups. No survey, no prompt. 45 questions, from 9 organizations.
The largest group was about rules
Of the questions we have written up, the biggest theme was governance and rollout: should the policy come first, who owns the output, do we tell clients. People weren’t asking what AI could do. They were asking what they were allowed to do with it.
13 were asked at more than one firm
The same worry, arrived at separately, by people who have never met. Those are the questions worth answering properly, and each has its own page in the answers library.
What it changed for us
- We lead with the rules when a firm is unsure, because that is what they are actually stuck on.
- We answer “can I trust it?” with a method for checking, never with a yes.
- We publish what each tool does with data, because nearly every trust question came back to that.
The counts, the themes and the limits of the data are in the research note.
