Posts
AI lead scoring trained on your actual closed-won deals
Generic scoring decays. Custom models trained on your win history sharpen every quarter. Here is the data plumbing that makes it work.
Wiring Claude Desktop to your CRM via MCP
A tactical playbook for connecting Claude Desktop to HubSpot or Salesforce through TexAu's MCP layer, with auth, scopes, and the first three commands worth running.
Tactical: building a 12-provider enrichment waterfall
A walkthrough of stacking Hunter, Snov, Apollo, RocketReach, ZoomInfo, Dropcontact, Clearbit, Lusha, and four more into a single cascade with stop-on-match logic.
The economics of waterfall enrichment
How a 12-provider waterfall pays for itself in the first thousand rows — and the per-field math that makes it inevitable.
Why every GTM team will run an MCP server by 2027
The Model Context Protocol won't be an AI feature. It will be the way GTM software talks to itself. Most teams are about three quarters away from realizing they need it.