· 5 min read
Most product frameworks now feel redundant
RICE, MoSCoW, the 2x2 you drew in a workshop — they were compression tools for scarce build capacity. AI just made that scarcity mostly go away.
Every product framework I've ever used was quietly solving the same problem: we could only build three things this quarter, so which three? Scoring models, prioritisation matrices, effort-versus-impact plots — all of them are rationing devices dressed up as thinking tools. They made the trade-off legible enough that a room full of people with different incentives could nod at the same roadmap.
That constraint has moved. When a working prototype takes an afternoon instead of a sprint, the cost of finding out is lower than the cost of arguing about it. Sitting in a two-hour session assigning confidence scores to a guess is now more expensive than just building the guess and watching what people do with it.
What the frameworks were actually doing
- Rationing scarce engineering time across too many stakeholders.
- Manufacturing consensus in rooms where nobody had real evidence.
- Giving a PM cover to say no without it feeling personal.
- Turning judgement into something that survived a steering committee.
Only one of those is about making a better product. The rest are organisational plumbing. Useful plumbing — but let's not confuse it with craft.
What's actually gotten harder
Deciding what to build is easier. Deciding what's true is harder. When you ship five variants in a week, your bottleneck stops being prioritisation and becomes evaluation: do you have a way of knowing which one is better, on a dimension your customers care about, before the market tells you the expensive way?
The scarce resource is no longer build capacity. It's the taste to know which of the ten things you just built is worth keeping.
That's why eval design, instrumentation, and a stubbornly specific point of view about the customer are eating the space frameworks used to occupy. A prioritisation score is a prediction. An eval is a measurement. When predictions get cheap to test, measurement wins.
What I'd keep
- Opportunity solution trees — because they force you to name the problem, not the feature.
- JTBD interviews — the raw material doesn't get cheaper just because building does.
- Anything that produces a written, falsifiable claim about a customer.
- A clear strategy narrative. Speed without direction is just more surface area to maintain.
What I'd retire
Scoring rituals where the numbers are invented to justify a decision already made. Quarterly roadmap theatre that can't absorb a result from Tuesday. Any artefact whose main audience is a slide, not a build.
The frameworks aren't wrong. They're just answering a question most teams no longer have. The new question is: now that you can build almost anything, what makes you the team worth building it?