discovery-synthesis

You can extract JTBD signals from archive.sqlite, customers.md, and memory ledger entries. This approach analyzes existing per-org data stores to identify problem patterns and knowledge gaps, delivering data-driven insights without requiring direct user interviews.

Yes, you can optionally incorporate interview transcripts to enhance signal quality. While the core analysis relies on archive.sqlite, customers.md, and memory ledger data, uploading transcripts enriches the extracted JTBD signals and problem patterns.

JTBD signals, problem patterns, and knowledge gaps are output as structured JSON. The JSON includes specific fields for jtbd_signals, problem_patterns, and gaps, along with evidence_refs and blocking data where applicable.

The best way is to analyze existing memory ledger entries and customers.md files for recurring problem signals. This process extracts JTBD signals and identifies knowledge gaps from your per-org data stores to guide product discovery efforts.

A memory ledger is part of the core data sources used alongside archive.sqlite and customers.md. These three per-org data stores are queried together to extract JTBD signals and problem patterns, ensuring comprehensive customer insights.