Problems

What the feedback adds up to, and your call on each

Updated 1 Nov 2026 · 3 min read

On this page
  1. How problems are ranked
  2. The views
  3. A problem's page
  4. New patterns
  5. Who can do what

Problems (in Learn) is what Mind found in your feedback: pains, in your users' words, each with the feedback behind it, ranked, and your call on each. It replaces the Intelligence Hub: every insight became a problem (P-1, P-2, …) with its feedback, the call you made on it and its old key. Old Hub links, insight links included, open the problem.

How problems are ranked

Every night Mind works out each problem's evidence strength from your users' feedback: how many say it (the last 90 days), how much of it comes from your main customers (ICP fit), who pays, the trend over two weeks, and any ranked survey you imported. The drivers show under the score: read them, they are not a formula to tune.

  • Your users come first. Reddit threads, competitor reviews and public threads you paste are shown ("Also in public: N · not counted") but never count. A problem with only public evidence is listed under Public signals, not ranked.
  • No invented rankings. Below about 30 pieces of feedback in the product, or with fewer than 3 pieces behind a problem, it says "Not ranked" and why.

The views

Ranked (open problems), Needs your call, New patterns (what Mind spotted; not a problem until you say so), Parked and closed, and All. Each row shows the feedback (all · this week), five weeks of it, ICP fit, the score, the status and Mind's suggestion. The eye opens a quick look; the name opens the problem.

A problem's page

  • Mind's read: what the feedback says, with its sources and date. Read it again asks Mind for a fresh read (1 AI credit). A read carried over from the Intelligence Hub says so.
  • Your call: Mind's suggestion first, then Build, Test first, Not now or No. Every call needs a reason in one line, and the dialog says exactly what will happen. Build makes a feature in Ideas (or joins one you pick); the problem's feedback shows on it as who asked. Not now asks you again on a review date (30 days by default). No keeps it closed; new feedback still lands on it. Reopen puts a closed problem back. Every call can be undone for 24 hours.
  • Who feels it: the people behind it by segment (your main customers first), with their own words behind each Why?. "Felt most by …" shows from 5 people on. If Mind doesn't know who your users are, tell it once on Problems.
  • In their words: quotes, exactly as written. Public ones are marked "Not your user".
  • By week, Where it came from, its tags, and Before Problems: everything the old Hub knew about it, kept as it was.
  • Feedback, People and History tabs: every piece, every person, and every call, merge, rename and new read with its date and who made it.

More lets editors Merge a problem into another (its feedback moves; both names stay findable) or Rename it.

Test first

When you decide a problem Test first, its page offers Draft the test: a draft test of that problem, from the idea you wrote. It opens on its own page for you to set what would make you believe it (the pass line) and the dates before it starts. No credits. Once it exists, the problem links it ("Its test: T-3").

New patterns

Mind proposes new patterns from feedback filed under no problem. On each: Make it a problem (ranked tonight), Keep as a tag under a problem it belongs to, or Leave it in the area.

Who can do what

Everyone on the product reads Problems. Editors and admins make the calls. Viewers see names as initials in quotes.