Analyzing Insights & Themes
Learn how evidence is automatically extracted from interviews and transformed into actionable insight themes
When you upload an interview (audio/video or transcript), our AI automatically processes it through multiple stages to extract meaningful evidence:
Audio/video is transcribed with speaker identification, timestamps, and turn-by-turn dialogue segmentation.
AI identifies verbatim quotes that contain valuable insights - customer pain points, needs, motivations, outcomes, workflows, and contexts.
Each piece of evidence is tagged with facets - structured attributes like pain points, jobs-to-be-done, desired outcomes, emotions, motivations, contexts, and workflows.
Evidence is automatically linked to the people and organizations mentioned, creating traceability from raw quotes to specific customers.
💡 Fully Automatic
This entire process happens automatically when you upload an interview. No manual tagging or coding required—just upload and let the AI extract the evidence.
Themes are strategic, high-level patterns that emerge from clustering similar evidence across multiple interviews. The system targets 3-7 major themes per project, not 30+ micro-themes.
Evidence with similar meanings is automatically grouped together using embeddings. The AI aggressively consolidates related pain points, workflow steps, and industry variations into strategic themes.
The AI merges symptoms into root causes, combines workflow steps, and focuses on business-decision driving patterns. Each theme represents 5-20+ pieces of evidence, not individual micro-issues.
AI generates clear, strategic names for each theme and automatically links them to people and organizations for segmentation analysis.
Click "Enrich Themes" to automatically add rich metadata including pain points, jobs-to-be-done, desired outcomes, emotional responses, and categories.
✨ When to Enrich Themes
Run theme enrichment after you've uploaded 5-10 interviews and themes have been auto-generated. Enrichment is computationally expensive, so you typically only need to run it once themes are stable, then refresh every month or when you add significant new evidence.
🎯 Theme Count Guidelines
Expect 3-7 strategic themes for most projects. If you see 20+ themes, the system is being too granular. Very large projects (100+ interviews across multiple product lines) might justify 8-12 themes, but rarely more.
When themes are enriched, they gain structured metadata that helps you quickly understand and act on customer insights:
Pain Point
The specific problem or frustration customers experience. Use this to prioritize fixes and improvements.
Job To Be Done
The underlying goal customers are trying to accomplish. Use this to design features that truly help customers succeed.
Desired Outcome
What success looks like for the customer. Use this to measure feature impact and validate solutions.
Emotional Response
How customers feel about the situation. Use this to understand urgency and prioritize emotional impact.
Every theme is automatically linked to the people and organizations that mentioned it, enabling powerful segmentation analysis:
📊 What You Can Analyze:
🔍 Where to Find Coverage Data:
💡 Example Insights
"5 healthcare companies mentioned CRM Efficiency issues, but only 2 tech startups did" → Healthcare vertical needs different CRM features. "CTOs care most about Reliability, VPs care about Analytics" → Tailor messaging by persona.
📝 Use Evidence When:
🎯 Use Themes When:
💡 Best Practice
Start with themes to identify high-level patterns and priorities. Then drill down into evidence when you need to validate findings, gather quotes for presentations, or understand the nuances of a specific theme.
Build a meaningful evidence base. Each interview is automatically transcribed, evidence extracted, and facets tagged.
Navigate to Themes page. The system automatically creates strategic themes from your evidence. New interviews link to existing themes or create new ones if truly novel.
Click "Enrich Themes" button to add pain points, jobs-to-be-done, desired outcomes, and emotional responses. Refresh monthly or when adding significant new evidence.
View theme detail pages to see which organizations and people mentioned each theme. Identify persona-specific or industry-specific patterns.
Click evidence links to see raw verbatim quotes, validate insights, and gather customer voice for presentations or stakeholder reports.
Keep uploading interviews as you conduct them. Evidence automatically extracts and links to existing themes. System stays self-organizing without manual theme management.
✅ Fully Automated After Initial Setup
After your first enrichment, the system runs hands-free. New interviews → Evidence extraction → Theme linking happens automatically. You only intervene to enrich new themes or when you want updated metadata after major evidence growth.