AI mode on the product maps
What the AI review on the Feature Tree and Product structure shows, how labels are decided and how many credits it uses.
Turn on AI mode on the Feature Tree or the Product structure. A review panel opens beside the map.
Feature Tree. Bonzaii calculates measures for every branch (each top-level feature, and each deeper feature with children): size, solutions planned, in progress and shipped, readiness, evidence, importance and impact, last change, and features with no screen. Each branch then gets one or more labels, compared with the other branches in this product:
- Focus here — important, but readiness is weak, little has shipped recently or open problems point at it.
- Under-developed — important, but few child features or solutions compared with similar branches.
- Over-developed — many features or a lot of recent investment, but low importance or impact, or no evidence.
- Unresearched — no linked insights, research or discoveries, or most features have none.
- Blocked — blocking tasks hold back a large share of its features.
- Stale — nothing changed or shipped for 6 months.
- Healthy — none of the above.
The panel shows a summary, the branches to focus on with a next step, readiness, a Balance section and a level view. Click a branch on the tree to see why it got its labels and what you can do: create a solution, add research or a discovery, open Readiness, and add a feature (or, in AI builder products, report a mapping issue).
Product structure. Lists screens with no features, features with no screen, journeys with gaps, and screens linked to blocked or unresearched features.
The AI only writes the wording; it can't add numbers or verdicts. If a branch lacks scores it says "Not enough data". Each review uses 2 credits, is charged only when it succeeds, and is kept for 24 hours — use refresh to run it again.