Resonance Sales Call Intelligence

/resonance-sales-call-intelligence: analyze sales transcripts

> Role: resonance-sales > Input: Sales call transcripts (pasted text, CSV, or MCP connections). > Output: A structured Persona Intelligence Report and Interactive React Dashboard. > Definition of Done: The output dashboard parses speaker turns, extracts goals, pains, triggers, objections, and product feature gaps with direct verbatims, and provides a messaging vocabulary library. Free of AI slop and em dashes. Passed the validator.

Prerequisites (fail fast)

• [ ] Transcripts or call recordings are available for analysis.
• [ ] You have identified if speaker attribution (rep vs. prospect) is present or needs auto-inference.

Algorithm

Copy this checklist and tick items as you go.

1. Profile Prospects: Extract job titles, seniority levels, and target department groupings (e.g., C-Suite, VP, Manager). → verify: dataset confidence level is declared at top.
2. Ingest and Clean: Ingest raw text or CSV entries, map speakers, and separate conversation blocks. → verify: transcript source type is declared.
3. Extract 10 Dimensions: Extract goals, pains (functional/emotional/social), triggers, objections, product gaps, competitor comments, buying processes, vocabulary words, buying signals, and red flags. → verify: all quotes are extracted exactly as spoken.
4. Objection & Feature Analysis: Tabulate objection types and rank product feature gaps by frequency. → verify: objection handling is rated as Effective, Neutral, or Missed.
5. Interactive Dashboard: Construct an interactive React-based dashboard displaying tabs for Overview, Personas, Objections, Feature Gaps, and Vocabulary. → verify: contains a bar chart representing objection frequency.

Recovery

• Transcript dataset size is extremely small (1-2 calls) → label confidence as LOW and append a standard directional hypothesis warning.
• Speaker attribution is missing or mixed → run speaker role auto-inference logic (detecting who explains pricing/product vs. who highlights constraints) before continuing.
• Tried to build a dashboard 3 times but React components hit rendering errors → stop, output the structured long-form markdown report instead, and escalate.

Operating Standard

Apply the Resonance operating standard from AGENTS.md (always loaded): the builder Voice and its banned-word list (no AI slop, no em dashes), Recommendation-First decisions (models recommend, the user decides), the Completion protocol (end with DONE / DONEWITHCONCERNS / BLOCKED / NEEDS_CONTEXT, backed by evidence, escalate after 3 failed tries), and the Ratchet (log durable learnings to .resonance/learnings.jsonl).

> Model note (Claude): Strong native reasoning. Do not narrate "let me think step by step" or pad with chain-of-thought; think, then act. Prefer the dedicated file and search tools over shell. State assumptions briefly, then proceed.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: manusco
• Source: manusco/resonance
• License: MIT

Install and usage instructions live in the source repository linked above.