Purpose
Transform unstructured notes and transcripts from executive customer conversations into a clustered, prioritized set of themes — each tagged by frequency, sentiment, urgency, and the function that owns the follow-up — so insights stop dying in meeting notes and start routing cleanly to Product and GTM. Solves the synthesis bottleneck that keeps a Customer Insights Lead from reporting patterns across dozens of conversations.
When to Use
Invoke this skill when the user asks to:
- Synthesize notes or transcripts from a customer advisory board (CAB) session
- Pull recurring themes out of multiple executive briefing or customer interview transcripts
- Cluster customer asks and feature requests for routing to Product
- Identify what GTM, Product, and Leadership each need to act on from VoC data
- Turn raw call notes into a structured Voice of Customer report
- Find what customers are repeatedly asking for across a quarter of conversations
Inputs Required
Before starting, collect:
- Conversation source(s): Raw notes, transcripts, or summaries — pasted or referenced. Multiple sources are expected and encouraged.
- Conversation context: What kind of conversations these are (CAB, exec briefing, win/loss, renewal QBR, discovery)
- Account context: Segment, industry, ARR tier, or strategic importance of the accounts represented (even loose labeling helps weighting)
- Routing targets: Which functions should receive insights (default: Product, GTM, Customer Success, Leadership)
- Strategic priorities: Current company bets or themes leadership is watching (so the synthesis can flag alignment or tension)
- Minimum theme threshold: How many mentions make a theme worth surfacing (default: appears in 2+ conversations OR raised by 1 high-priority account)
If account context or strategic priorities are missing, proceed with the conversations as-is and note where account weighting would sharpen prioritization.
Instructions
Step 1: Extract Raw Signals
Read every conversation source and pull discrete signals before clustering. For each signal, capture:
- The signal: A single ask, complaint, priority, or observation — one idea per signal
- Source: Which conversation/account it came from
- Verbatim: The closest direct quote available (preserve the customer's actual language)
- Type: Ask / Pain / Priority / Praise / Risk
Do not cluster yet. The goal of this step is exhaustive capture — surface every distinct point, including one-offs.
Step 2: Cluster Into Themes
Group signals into themes by underlying intent, not surface keywords. A theme is a recurring need or pattern, not a feature name. Apply these clustering rules:
- Merge signals that express the same underlying need even when worded differently ("the dashboard is slow" + "I can't get reports fast enough" → "Reporting performance and speed")
- Keep distinct needs separate even when they touch the same product area
- A theme qualifies for the report if it meets the minimum theme threshold; below-threshold signals go to an "Emerging / Watch" list
For each qualifying theme, compute:
- Frequency: Number of conversations the theme appeared in, and total signal count
- Sentiment: Positive / Neutral / Negative / Mixed
- Account weight: Whether high-priority or strategic accounts raised it
- Urgency: Now (blocking/churn risk) / This quarter / Directional
Step 3: Prioritize Themes
Rank themes using a combined lens, in this order of weight:
- Revenue/retention risk — themes tied to churn, renewal, or expansion blockers
- Frequency × account weight — broadly raised, especially by strategic accounts
- Strategic alignment — themes that reinforce or contradict current company bets
- Effort-to-impact signal — where the customer language suggests a quick win
Produce a ranked list of the top themes. Be decisive about ranking — do not present them as equally weighted.
Step 4: Route Each Theme
Assign every ranked theme to an owning function and define the handoff:
- Product: Feature gaps, roadmap asks, usability/performance issues, integration needs
- GTM (Sales/Marketing): Positioning gaps, competitive mentions, messaging misalignment, pricing/packaging friction
- Customer Success: Onboarding, adoption, support, enablement, relationship risks
- Leadership: Strategic tensions, market shifts, themes that cut across functions or challenge a current bet
For each theme, write a one-line routing note: what the owning function should do with it and why it matters now.
Step 5: Build the Routing Table and Executive Summary
Compile the synthesis into the output structure below, leading with an executive summary that a leader could read in 60 seconds.
Output Format
- Structure: Executive summary → ranked theme cards → routing table → emerging/watch list → recommended next actions
- Length: 1,000–1,500 words depending on conversation volume
- Format: Structured document — executive summary in prose; theme cards as labeled blocks; routing as a table
- Sections: 1. Executive Summary (4–6 sentences): top 3 themes, the single biggest risk, and the one thing leadership should act on 2. Ranked Themes (theme cards): each with name, one-line description, frequency, sentiment, urgency, account weight, 1–2 supporting verbatim quotes, owning function, and routing note 3. Routing Table: columns — Theme | Owner | Urgency | Action | Source Accounts 4. Emerging / Watch List: below-threshold signals worth monitoring 5. Recommended Next Actions: 3–5 concrete follow-ups with owners
- Tone: Customer insights lead briefing executives — analytical, decisive, evidence-backed. Let customer verbatims carry the emotional weight; keep the framing neutral.
Quality Checklist
Before delivering output, verify:
- [ ] Every theme card shows a frequency count (conversations + signals) — no theme is asserted without evidence
- [ ] Each theme carries at least one verbatim customer quote drawn from the source material — not paraphrased into corporate language
- [ ] Themes are clustered by underlying need, not by surface keyword (no two cards describe the same need)
- [ ] Every ranked theme is routed to exactly one owning function with a specific action, not "share with the team"
- [ ] The routing table accounts for every ranked theme — no theme appears in the cards but is missing from the table
- [ ] The executive summary names a specific biggest risk and a single recommended action — it does not hedge
Notes
- The most common synthesis failure is keyword clustering — three customers describing the same frustration in different words must land in one theme, or the frequency count understates the signal.
- Preserve customer verbatims aggressively. A direct quote from a strategic account moves internal teams more than a clean summary ever will.
- When a theme contradicts a current company bet, surface the tension explicitly in the Leadership routing rather than softening it — that contradiction is often the most valuable output of the synthesis.
- If conversations span a time range, note whether a theme is rising, steady, or fading; trajectory often matters more than raw count for prioritization.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills