Using MCP for CRM data cleanup: auditing 14k stale HubSpot accounts instead of prospecting
snowingbol · reddit · 2026-09-04
Instead of the common GTM use case of finding prospects, the author's team uses MCP to find bad data in HubSpot: 14,000 accounts accumulated over six years, full of dead companies, outdated employee counts, and long-gone contacts.
Stack: HubSpot is the thing being checked, Slack receives weekly diffs, Notion holds the correction log, and Coresignal serves as the external reference for company/employee data.
Key insight: the useful queries are comparisons, not searches — accounts with wildly off employee counts, companies with no apparent current employees, and contacts who've left their attached company (the biggest headache-saver, surfacing former champions still sitting in the CRM).
Pitfall: initially the external records landed via monthly reloads, so they were comparing stale CRM data against a stale reference — employment changes now arrive as events instead. Everything is read-only; the agent flags problems but never touches records without human review.
More from coding & agent
- Sakana AI's Stefania Druga demos agentic memory experiments for edge devices — kaixhin · 2026-09-04
- rybbit-mcp lets Claude Code query Rybbit Analytics via natural language — modelcontextprotocol · 2026-09-04
- Agentic Coding's Biggest Flaw: Managing AI's Pointless Code Rewrites — kylegawley · 2026-09-04
- Anthropic reveals 3 Claude sandbox escapes, one touched a production database — Sumsub_Insights · 2026-09-04
- Subagents when the API key hits the rate limit — realsohamparekh · 2026-09-04
- Yoav Goldberg pushes back on DHH: an agent-built Qt markdown editor isn't a path to personal Photoshop — yoavgo · 2026-09-04