Practical field notes on finding the evidence your CRM, health score, and forecast leave out.
AI Agents·MCP·Context·RevOps
SaaS spent 20 years selling silos. Agents and MCP finally give us a way out, and the incumbents are pulling everything back into their own box. The silo was never the database. It was the meaning.
Craig Tracey ·
Agents·Context·Governance·MCP
Support agents fail in production because they lack live context: who owns the account, what they use, what is at risk. Live graphs fix the substrate.
Craig Tracey ·
MCP·Agents
Getting an MCP server working is not the hard part. Operating one is: who can access what, catching failures early, and enough visibility to improve.
Craig Tracey ·
MCP·Agents
MCP has moved fast, but implementation quality lags adoption. Tool design and context quality are the two areas where most MCP servers fail.
Craig Tracey ·
Agents·MCP
We benchmarked six LLMs with 25 to 150 MCP tools and measured accuracy loss, latency spikes, and hard API limits. The cheapest model won.
Craig Tracey ·
MCP
The debate between MCP and CLI for agent tooling misses the point. The real question is which mode you are building for, and where token costs hide.
Craig Tracey ·
MCP·Agents·Architecture
Giving an LLM 40 tools and hoping it picks the right one repeats the mistake dashboards made for humans. For agents, the fix works differently.
Craig Tracey ·
MCP
MCP gives LLMs powerful tools, but without relationship data every request becomes trial and error. That is where hallucinations come from.
Craig Tracey ·
AI·MCP·Platform Engineering
Introducing SixDegree: live operational state for the agentic enterprise, auto-derived from your systems of record and served to AI agents via MCP.
Craig Tracey ·