Duo Copilot Dismiss Diagnostics

Help admins understand what's driving their Duo dismiss rate by analyzing dismiss patterns across reasons, signals, lead attributes, and rep behavior.

Instructions

When an admin wants to diagnose their Duo dismiss rate, query analytics and analyze the results top-down to find the biggest contributors and recommend fixes.

Steps

1. Confirm scope. Default to last 30 days, full account. Ask if the user wants a different timeframe or team scope.

2. Gather the data by submitting these analytics questions via mcp__claude_ai_Amplemarket__ask_analytics:

   • Overall dismiss rate and lead counts (generated, actioned, dismissed)
   • Dismiss reason breakdown (Bad lead, Bad company, Bad signal, Other)
   • Dismiss rate by signal type and signal name
   • Types of companies being dismissed (sizes, industries, etc.)
   • Profiles of dismissed leads (seniorities, titles, etc.)
   • Dismiss rate by rep

   Submit all in parallel, poll with mcp__claude_ai_Amplemarket__get_analytics_result.

3. Analyze top-down. Work through the data in this order:

   1. Reasons -- Which dismiss reason dominates?
   2. Patterns within that reason -- What's common about the dismissed leads or companies?
   3. Signal contribution -- Which signals drive the most dismisses in absolute terms?
   4. Rep patterns -- Is it universal or concentrated in a few reps?

4. Present a well-formatted diagnostic report. Lead with a summary headline, then use tables and clear sections to walk the admin through the findings. End with specific, actionable recommendations tied to Duo configuration changes.

Examples

User prompt: "Our Duo dismiss rate is at 38%. What's going on?"

Simple example output: The agent finds that Bad company is the top dismiss reason (56%), discovers most of those are sub-50-employee companies coming from the Hiring Surge signal, notes rep dismiss rates are uniform (so it's a signal issue, not a rep issue), and recommends excluding small companies from Hiring Surge -- projecting the dismiss rate would drop to ~22%. All presented in a structured report with tables for the reason breakdown, signal contribution, and rep patterns.

Troubleshooting