Better decisions. More useful products.

Next edition · Draft for review. Adapted from six original articles by Raghvendra Singh. This edition has not been sent or scheduled.

Start with the decision

Across my writing on products, AI, leadership and teaching, one question keeps returning: what decision are we helping someone make? A polished screen, fluent AI answer or busy review calendar can still leave that question unanswered.

A method for your next review

Choose one product moment. Write the decision someone must make, the information they need, the doubt that could block them, the next action and the outcome they should be able to recognise. Use that stack to review the work before discussing visual polish.

For founders and team leaders

Make decision ownership visible. Record who decides, what evidence they use, what is waiting and what happens when the decision is wrong. In critique, give the team a written quality bar and separate personal preference from risk.

For designers and AI practitioners

Make the reasoning inspectable. In an AI workflow, identify what the system can do alone, when it must ask and where a person signs off. Show sources, uncertainty and an undo path. In your own practice, explain what you kept, what you cut and which evidence changed your decision.

An example to think with

My article uses a lending offer as an illustrative example: the problem may be uncertainty about repayment, fees or tenure rather than the layout of the offer card. This is a method example, not a claim about measured conversion improvements.

Try this before your next meeting

Pick one important screen or workflow. Write its decision in one sentence. List three doubts that could stop the user. Check which doubts the experience actually answers. Bring that list to the review, then choose one change to test.