“We want an AI agent.” It sounds like a brief. It is not one.
An agent is a tool. The requirement has to describe the outcome: which work disappears? which decision gets faster? which risk must the machine not carry?
Without that, you build theatre. A chat window that impresses, and that nobody opens after six weeks. The agent was never the goal. The goal was a change in daily work.
My test is simple: can the team measure success in one sentence without using the word AI? If not, the requirement is not ready.
Written by me. The thoughts, the values, the learnings, the mistakes: all mine. Grammar and spelling are corrected by a twin model of my own, trained on my texts. Sometimes that turns into a mix. Of course I read everything again, but nobody is infallible. Whatever stays in is mine too.