Turn sales calls into a living buyer persona database

Before routing any transcripts, define the fields PMM actually needs to analyze. Create the database in Notion or Airtable with fields for buyer role, segment, use case, pain, trigger, current workaround, buying committee role, decision criteria, objections, competitor mentions, exact buyer language, evidence quote, call URL, account name, opportunity stage, confidence score, and review status. Add allowed values where possible so the system does not produce twenty versions of the same persona label. Create views for unreviewed records, approved records, persona themes, and monthly changes. This schema is the difference between a useful research system and a pile of AI summaries.

Design a persona intelligence database schema for PMM.

ICP:
{{icp}}

Current persona hypotheses:
{{persona_hypotheses}}

Sales motion:
{{sales_motion}}

Products or use cases sold:
{{product_use_cases}}

Return:
1. Recommended database fields
2. Field type for each field
3. Allowed values where useful
4. Required versus optional fields
5. Evidence rules
6. Confidence scoring rules
7. Review statuses
8. Example filled record
9. Views PMM should create for analysis

Keep the schema practical enough to maintain every month.

Use Gong, Salesforce, and Zapier to route only the calls that can improve persona intelligence. Include discovery calls, demo calls, evaluation calls, late-stage buyer meetings, and win-loss calls with enough buyer participation. Exclude internal meetings, short scheduling calls, customer support calls, implementation calls, and calls where the transcript is mostly the seller talking. Add filters for opportunity stage, call duration, account segment, call type, and whether a transcript is available. Create a rejected-call log so you can later inspect whether the filters are too strict or too loose.