skills.rest
amplitude
amplitudeOFFICIAL
Explore all skills by this author →
WebsiteGitHub

analyze-ai-topics

Analyzes AI agent conversation topics to surface underserved areas and coverage gaps.

Updated Apr 3, 2026
Search Tags:#ai agents#product insights#topic analysis#amplitude#agent quality#coverage gaps#conversation analytics
One-click install
npx skills add https://github.com/amplitude/amplitude-copilot-plugin --skill analyze-ai-topics-amplitude
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: analyze-ai-topics
Source: https://github.com/amplitude/amplitude-copilot-plugin/tree/main/skills/plugins/amplitude/analyze-ai-topics
Command: npx skills add https://github.com/amplitude/amplitude-copilot-plugin --skill analyze-ai-topics-amplitude

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Product teams running AI agents lack visibility into what users actually ask and where the agents underperform, making it hard to prioritize improvements from raw conversation data. ## Core Features & Use Cases - Topic Landscape Mapping: Breaks down conversation volume, quality scores, sentiment, and failure rates by topic across all agents. - Underserved Topic Detection: Scores topics on a volume-by-quality matrix to find high-traffic areas where agents perform poorly. - Coverage Gap and Routing Analysis: Identifies questions agents cannot answer, emerging topics, and topics better served by a different agent. - Use Case: Ask "what are people asking our AI about?" and receive a prioritized topic heatmap with conversation excerpts, failure analysis, and concrete recommendations for what to fix first. ## Quick Start Ask the assistant to analyze what users are asking your AI agents and where the AI is struggling.

Frequently Asked Questions about analyze-ai-topics

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

FAQPage Schema
How do I find out what users are asking my AI agent?▼

Use Amplitude Agent Analytics topic breakdowns to see each conversation topic with session counts, quality scores, and sentiment. This skill queries the topic metrics and presents a heatmap sorted by priority so you can see the full landscape at a glance.

How to identify where an AI agent is underperforming?▼

Score topics on a volume-by-quality matrix: high-volume topics with quality scores below 0.7 are underserved and need fixes first. Supplement with failure-rate data and sampled conversations to understand exactly where the agent goes wrong.

Does this analysis require Amplitude Agent Analytics instrumentation?▼

Yes, the skill only works when Amplitude Agent Analytics is instrumented in your project, since it relies on topic enrichment, session metrics, and conversation search from the Amplitude MCP tools. Without instrumentation, no topic data is available.

What if topic data is empty or too generic?▼

Empty topics usually mean session enrichment is not enabled; the skill falls back to keyword-based conversation searches to manually categorize themes. If labels are too broad, it notes the enrichment model may need tuning and derives sub-topics from conversations.

Can I analyze topics for just one specific agent?▼

Yes, you can scope the analysis to a single agent, time window, or focus area. The skill then compares that agent's per-topic quality scores against the fleet average and deep-dives into its weakest topics.

Related Skills

Explore More Product & Management Skills→
amplitudeamplitude
28

integration-javascript_web

Integrate Amplitude analytics into client-side JavaScript web applications.

Official
Intermediate
monkey1saimonkey1sai

saga-observability-platform

Track AI agent reasoning traces and detect inference anomalies in real time.

Community
Advanced
citedycitedy
31

Feedback Synthesizer

Collect and analyze user feedback from multiple channels into actionable product insights.

Official
Advanced
manojbajaj95manojbajaj95
83

data-and-funnel-analytics

Set up GA4/GTM tracking and analyze conversion funnels and ROI.

Community
Advanced