What this agent does
Stop letting inbound leads sit before triage. This agent runs a two-stage pipeline that researches, scores, and drafts a response for every lead as it arrives.
stage-1-scoring— researches the company and the person's role from public sources, scores the lead against your Ideal Customer Profile, and assigns a Hot/Warm/Cold tier with a routing recommendation. Never fabricates a company fact, role, or signal.stage-2-draft-reply— drafts a tailored first-touch reply matched to the lead's tier and researched context, and logs the lead into a local tracker. Never sends anything on its own.
Hot leads get a specific personal draft; cold leads get a brief generic one — every lead is logged for review and approval.
Requirements
- Runs entirely on free, built-in research tools (
web_research,web_search,maps_search) — no paid lead-enrichment, firmographic, or people-search API required. - Works on Zo Computer natively and on Claude AI (claude.ai with web search, or Claude Code).
- To actually send the first-touch reply (a separate, manual step you control), an email integration needs to be connected — this project only ever produces the draft.
Setup
- Unzip the package into your workspace, keeping the folder structure intact.
- Fill in your real Ideal Customer Profile (industry, size, authority signals, disqualifiers) — the default placeholder produces only provisional scores.
- Run a starter prompt with a real lead's name, email, and company to see the full pipeline end to end.
What you get
- Every inbound lead automatically researched and scored against your Ideal Customer Profile, with no fabricated facts
- A Hot/Warm/Cold tier and routing recommendation for each lead the moment it arrives
- A tailored first-touch reply draft matched to the lead's tier and research, ready for your review before sending
- A local lead tracker logging every lead for approval and follow-up
- Setup that works out of the box on both Zo Computer and Claude AI, no paid enrichment API required