stage2-extraction

Extract patterns, FAQs, and response strategies from clustered customer support transcripts.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/channel-io/sop-agent --skill stage2-extraction
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: stage2-extraction
Source: https://github.com/channel-io/sop-agent/tree/main/.agents/skills/stage2-extraction
Command: npx skills add https://github.com/channel-io/sop-agent --skill stage2-extraction

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pandas, openpyxl, and includes scripts (resource) components.

What problem does it solve?

This SOP enables data teams to derive concrete patterns, FAQs, and response strategies from real chat transcripts after Stage 1 clustering, reducing manual analysis workload and improving SOP accuracy.

Core Features & Use Cases

  • Enriches Stage 1 results by reading full conversation transcripts (turns) rather than summaries to identify 3-8 patterns per cluster.
  • Produces a SOP topic map that maps clusters to authoritative topics and HT/TS classifications, enabling streamlined Stage 3 generation.
  • Generates FAQ pairs and response strategies grounded in verbatim customer phrases for realistic, channel-appropriate guidance.

Quick Start

Run the enrichment workflow to extract full conversations and patterns from Stage 1 results.

Frequently Asked Questions about stage2-extraction

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract SOP patterns and FAQs from clustered customer support conversations?

To extract SOP patterns and FAQs from clustered conversations, you enrich Stage 1 cluster results by reading full chat transcripts to identify 3-8 verbatim patterns per cluster. This process outputs structured patterns.json and faq.json files for downstream automation.

What is the best way to map customer support chat clusters to authoritative topics?

The best way to map chat clusters to authoritative topics is to analyze full conversation turns rather than summaries, assigning standardized topics and HT/TS classifications to correct mislabels and enable streamlined Stage 3 SOP generation.

Can I generate response strategies from raw chat transcripts using pandas?

Yes, you can generate response strategies from raw chat transcripts using pandas. This Skill leverages pandas and openpyxl to process conversation data and output response_strategies.json grounded in verbatim customer phrases for channel-appropriate guidance.

Do I need Stage 1 cluster results before extracting conversation patterns and topics?

Yes, you need Stage 1 cluster results before extracting conversation patterns. This Skill enriches Stage 1 outputs by ingesting full conversation transcripts to discover patterns, assign topics, and correct mislabels for subsequent SOP creation.

What specific outputs do I get when converting clustered chats into SOP patterns?

Converting clustered chats into SOP patterns yields four JSON outputs: patterns.json for conversation trends, faq.json for question pairs, response_strategies.json for guidance, and keywords.json for downstream enrichment and automation workflows.