Product and portfolio management
Rebuilding product decisions around AI evidence.
Portfolio decisions have always been evidence-starved. The data arrives late, the signals are noisy, and by the time the quarterly review has the numbers, the market has moved. Product teams compensate with judgement, which is a polite word for well-informed guessing.
AI shifts the evidence base, modestly but usefully. Usage telemetry, support signal, pricing elasticity, competitive moves, and buyer language can now be synthesised continuously rather than quarterly. The portfolio conversation can be held against evidence that is weeks old rather than months old.
This does not remove judgement. It raises the bar on it. When the evidence is fresher and more granular, the executives in the room have to justify decisions against what the data actually says, not against the data they wish they had. That is uncomfortable for product leaders used to winning rooms on narrative.
The predictable failure mode is building the evidence pipeline and then ignoring it when it says something inconvenient — kill this product, deprioritise that bet, the beloved feature is not moving the number. If the governance does not let the evidence overrule the sponsor, the pipeline is theatre.
The AI is the easy part. Letting it inform the decision is the hard part.
The decision: before building the evidence pipeline, agree how the portfolio committee will handle evidence that contradicts the sponsor.