AI product management
Use AI to move faster without skipping problem framing, customer evidence, or evaluation.
Practical guides for founders, first PMs, product engineers, and product builders: discover customer problems, decide what to build, and ship with AI.
Field guides for small teams building software, AI products, and better product habits.
Use AI to move faster without skipping problem framing, customer evidence, or evaluation.
Go from a real customer problem to an AI feature you can evaluate, release carefully, and improve.
Practical workflows for people close to the code: learn from users, choose a small bet, and give coding agents useful context.
Turn customer signals into a focused plan, and test whether people value what your team is building.
A practical AI evaluation framework for product teams: build representative test sets, define a clear rubric, catch regressions, and connect model quality to user outcomes.
Practical guides, product frameworks, and Specky updates — sent when it’s worth your time.
The widely quoted 95% figure is not a universal AI failure rate. Here’s what the 2025 MIT NANDA report measured, what it leaves out, and how product teams can evaluate AI work against real outcomes.