There is a moment in almost every product meeting this year. The roadmap is on the screen. Someone has just finished explaining a perfectly reasonable feature. And then, from the end of the table, usually from someone who has been quiet until now: “Can we add AI?” Nobody laughs, because it is not a joke. Nobody objects, because objecting sounds like being against the future. So the question goes into the notes, gets a little sparkle icon, and quietly becomes a priority. I have heard it in boardrooms, in startups, and once, memorably, about a form with three fields.
“Can we add AI?” is not a feature request. It is a question about being left behind.
That is not a criticism. It is a reasonable fear. Everyone’s competitor has announced something. Every tool now has an assistant in the corner, waving. The person at the end of the table has read the same articles you have, and has to answer to someone who has read them too. But the question has the shape of a solution and the content of a feeling. It is not quite a strategy. Not quite a requirement. Not quite a user need. It is a mood with a budget attached. In my workshops with teams, I no longer argue with it. Arguing makes me the person against the future, and I have enough reputational problems as it is. Instead, I ask the question to finish its own sentence. Can we add AI so that who can do what, that they cannot do today? What happens next is very instructive. Sometimes the answer comes immediately, and it is good: a support team buried under the same twenty questions, a search that fails because people do not know the right words. In those cases, the question was never really about AI. It was about a problem, and AI happened to be the nearest tool. More often, the room goes quiet. Not the quiet of people thinking. The quiet of people realising that nobody in the room knows. That silence is not a failure. It is the most honest moment of the meeting, and the first one in which the actual design work can start. The useful version of the question was never “can we add AI?” It was “what is hard for our users right now?” Sometimes the answer needs a model. Often it needs a better label.
Margin note: Next time someone asks “can we add AI?”, don’t answer. Ask them to finish the sentence, and then wait. The waiting is the exercise. I’ve turned the sentence test into a one-page worksheet for product teams: the sentence, the “remove the word” check, and a space to sketch what happens when the AI gets it wrong. Download it here →
If your roadmap already has a sparkle icon on it, this is also where the Digital Product Design workshop begins. Let’s talk →
