Member-only story
Is AI the right solution? A 5-gate framework to help decide.
Use this 5-gate framework to evaluate if AI is the right solution for your product. Learn how to weigh the risks, costs, and benefits of AI against simpler alternatives.
Right, friends. It’s sprint planning on a Wednesday afternoon. A ticket lands from the Product Lead: “Add AI to onboarding to make it smarter.” No clear job to be done, no metric for what “smarter” means — or why we need it in the first place — just a loose promise that AI will make the thing “better”. You can almost hear the demo in someone’s head.
We’ve all been in that meeting. The pressure to “do something with AI” is everywhere, from board decks, OKRs, competitor checklists to LinkedIn and Medium (sorry to be adding to that, awkward).
It feels pretty inevitable, but the issue is, it’s a technology-first brief. It flips product development on its head: start with the tool, then go hunting for a problem. Normally, we do the opposite. We define the user job, the outcome, and the constraints, then pick the simplest approach that works.
That matters because AI carries real trade-offs: variable latency, a cost per request, probabilistic outputs (often persuasive, sometimes wrong), and operational needs like evaluation, monitoring, and on-call. Sometimes those trade-offs unlock serious value. Often…
