Make every discovery call summary double as a qualification scorecard

Retrieve the Discovery meeting type from the project's meeting taxonomy. Confirm it exists and has notetaker autojoin enabled — both are prerequisites for the custom summary recipe to fire on real meetings. If the type doesn't exist, halt and prompt the user to run /meeting-types-taxonomy first.

Configure the AI summary recipe for the Discovery meeting type. Extract structured fields: (1) Champion — name, title, level of buy-in (high/medium/low/none), explicit quotes showing commitment; (2) Pain — pain statement in prospect's own words, severity (must-solve/should-solve/nice-to-have), business impact discussed (revenue, cost, time, risk); (3) Current Solution — what they use today (vendor name + version), what works, what doesn't; (4) Decision Criteria — explicit criteria mentioned (price, features, integration, security, etc.), priority ranking if discussed; (5) Timeline — target go-live, urgency drivers; (6) Budget — explicit number, range, or signal (no budget signal = flag); (7) Next Step — what was agreed, with owner and due date. If a field is not discussed, return 'not_discussed' rather than guessing.

Compose a Discovery call quality dashboard reading from the structured summaries: % of discovery calls with all 7 fields captured (target: 70%+); breakdown of most-frequently-missed fields (signals coaching opportunities); discoveries with 'no budget signal' flagged for follow-up; champion strength distribution across recent discoveries; pain severity distribution. Group by rep so managers can spot reps consistently missing qualification fields.