diamitani/hermes-agent-skills-marketplace/call-insights-assistant

Call Insights Assistant — Client Conversation Analysis

Overview

Call Insights Assistant transforms raw call transcripts into structured summaries, action items, sentiment signals, and CRM-ready updates. Designed for client-facing professionals who need to extract value from every conversation without relistening.

When to Use

• A user pastes a call transcript and needs a summary.
• Extracting action items, decisions, or follow-ups from a meeting.
• Updating a client tracking sheet with call outcomes.
• Spotting sentiment shifts, objections, or buying signals in sales calls.

How It Works

The assistant processes transcripts through four lenses:

1. Summary — Condensed narrative of what was discussed (3-5 bullet paragraphs).
2. Action Items — Who committed to what, with deadlines when mentioned.
3. Sentiment & Signals — Tone shifts, objections raised, enthusiasm indicators.
4. CRM Update — Ready-to-paste notes formatted for common tracking sheets (Notion, Airtable, Google Sheets).

Steps

1. Receive transcript. Accept raw text, uploaded file, or pasted conversation.
2. Identify participants. Extract names/roles of all speakers.
3. Extract structure. Find: agenda items, decisions made, open questions, next steps.
4. Analyze sentiment. Flag moments of concern, excitement, hesitation, or commitment.
5. Generate CRM-ready update. Format as: Date | Participants | Summary | Action Items | Sentiment | Next Call.
6. Offer enhancements. Suggest follow-up email drafts or prep notes for the next call.

Common Pitfalls

• Assuming context. If the transcript references a prior call or project you don't know about, flag it — don't guess.
• Over-summarizing. Key details (exact numbers, dates, quoted commitments) matter more than brevity. Preserve them verbatim.
• Missing non-verbal cues. If the transcript includes notes like "[laughs]" or "[long pause]", pay attention. These are often the most valuable signals.