Dylan Ottinger

gtm-mcp-server - A single MCP server that makes the full GTM stack agent-accessible.

Wires HubSpot, Clay, Apollo, Slack, and email into one Model Context Protocol server so any MCP-compatible AI client (Claude, Cursor, etc.) can prospect, enrich, update CRM records, and fire notifications from a single prompt. A workflow that normally spans five tools and manual context-switching becomes one instruction: find VP Sales targets, enrich them, create contacts in HubSpot, and notify the team in Slack.

Why it matters: This gives agents a shared execution layer across the systems revenue teams already run on, so they can execute GTM workflows directly instead of just answering questions about them.

crm-reliability-engine - A human-in-the-loop CRM governance layer for AI-extracted GTM data.

Instead of letting Gong, HubSpot, or transcript-derived signals blindly overwrite CRM fields, the system turns every suggested change into a confidence-scored update proposal. Operators can approve or reject changes and every decision is logged in an event-sourced audit log.

Why it matters: Solves the "garbage in, garbage out" problem in automated CRM syncing, preventing data drift without requiring hours of manual sales rep auditing.

revenue-ops-system - Real-time inbound lead routing infrastructure for GTM teams.

AI enrichment adds intent tags and urgency signals, but routing stays deterministic: territory rules, rep capacity, and weighted round-robin logic decide ownership. Failed routes go into an exceptions queue and can be replayed by an operator.

Why it matters: Replaces fragile, black-box routing setups with audit-logged infrastructure so every lead's routing decision is traceable.