Customer Interview Synthesis

Customer interviews clustered into themes with severity, verbatim evidence, and a persona × theme heatmap - so the next product conversation starts with the four themes that matter, not 42 transcripts in a Drive folder.

What the managed workflow does

Clusters every transcript

Reads every interview, segments by JTBD, and tags each cluster against your prior-cycle theme dictionary so the diff is the story, not just this cycle's snapshot.

Scores severity

Mention frequency, persona spread, and product-usage correlation tag each cluster high · medium · low - so PM knows which 4 themes need investigation vs. background noise.

Surfaces verbatim

Anonymized direct quotes evidence each theme - the raw voice, not a synthesized paraphrase that strips away the texture PM teams need to actually act.

Cross-tabs by persona

A persona × theme heatmap shows where each issue concentrates - so the roadmap doesn't optimize for one persona's pain at the expense of another's job-to-be-done.

Themes clustered · severity-scored · evidenced with verbatim · cross-tabbed by persona

42 INTERVIEWS · 9 THEMES · 3 SEGMENTS

Inputs synthesized

Interview transcripts · 42JTBD framing · prior cyclesPersona segment · CRM tagsProduct usage signal

What lands per cycle

High-severity themesMedium · watchLow · monitor for driftAnonymized verbatim

Inputs synthesized

Interview transcripts. Every recorded session, transcribed, parsed, and clustered against your JTBD framing - not a PM's spot-read of 8 quotes from yesterday.
Prior-cycle theme dictionary. Themes carried forward so the diff is visible - what shifted, what disappeared, what's new this quarter.
Persona + segment metadata. Role, company stage, ARR band, tenure as a customer - drives the cross-tab without losing anonymity.
Product-usage signal. What each interviewee actually does in-product - so claimed pain is checked against observed behavior.

What lands per cycle

Theme tree. 3 trunks · 12 sub-themes · severity-scored, ordered by mention count, ready to share with PM and leadership.
Verbatim evidence. Anonymized quote cards per high-severity theme - so the next roadmap conversation starts with raw voice.
Persona × theme heatmap. Where each theme concentrates so the next bet matches the persona it's for, not a generic average.
Action prompts. Per high-severity theme: a calibrated next step - discovery sprint, prototype, pricing test - drawn from prior-cycle action library.

What you get, every week

An action-shaped readout

PM sees four themes with named actions and severity - not a 60-page Notion page of every transcript.

A diff against last cycle

Themes carry forward so movement is visible. The cycle becomes a continuous track, not a one-shot synthesis that drifts off relevance the day it ships.