.webp)
Table of Contents
Key takeaways
- Agent Connectors let Pigment's agents connect directly to external systems via MCP, the standard for AI tools reaching outside their own walls
- That means use cases like the Modeler Agent building a model from a spec in Google Drive, or the Analyst Agent flagging where performance is tracking against OKRs in Notion
- Every connection runs through one governed path between Pigment's agents and the systems around them, grounded in the planning data underneath
The bar for what AI agents are expected to do is rising constantly.
But context is king, and an agent can only reason with what’s in reach. For planning, the context that defines whether a model works often sits outside the platform:
- The requirements doc that defines what a model should do
- The assumptions page that explains why the numbers look the way they do
- The OKRs that you’re working toward
Today, teams bridge that gap by hand. Someone opens the spec, copies the relevant section, pastes it into a prompt, and hopes they caught everything.
Agent Connectors remove the manual steps. Pigment’s agents can now reach external systems through MCP, a common standard already used by tools like Claude, Cursor, and ChatGPT, without a bespoke integration for each one.
What this means for you
Connect the systems where your context lives, and the Modeler and Analyst Agents can read from them.
Context gathering
Let the Modeler Agent capture assumptions and requirements directly from the source:
- "Build the approved headcount reqs from this Google Doc into the workforce plan.”
- "Pull the supplier lead-time constraints from Notion and apply them to the inventory model.”
Cross-system reporting
Enable the Analyst Agent to access live business context:
- "Compare our latest forecast to the OKRs in Notion and flag where we're drifting."
- "Summarize this week's product performance from Pigment, combine it with the feedback in Slack"
“Notion brings teams, AI and shared knowledge together to move work forward. When Pigment’s agents can draw on that context, teams can ground planning and decision-making in the thinking they have already captured.”
Andrew McCarthy, General Manager EMEA, Notion
Want to see more?
You can, at the next Pigment live tour. Sign up for the next one.
To learn more visit our MCP page.
Frequently Asked Questions
What is Model Context Protocol (MCP)?
MCP is an open standard that gives AI agents a consistent way to connect to external tools and data sources. Instead of a custom integration for every system, an agent can reach any MCP-compatible source through the same protocol.
Which systems can Pigment's agents connect to?
Agent Connectors currently support Notion. Support for Google Drive and Slack will arrive in the coming weeks, with more connectors to follow.
Which agents use Agent Connectors?
Both the Modeler Agent and the Analyst Agent can pull external context through Agent Connectors as part of their workflows.
How is this different from the Pigment MCP Server?
The MCP Server lets external AI tools read governed Pigment data. Agent Connectors work the other way: they let Pigment's own agents read context from external systems. One brings Pigment data out to your AI tools; the other brings your business context into Pigment's agents.
Is the data governed?
Yes. Agents inherit the permissions of the person using them, in Pigment and in the connected system. An agent can only reach what that person could reach themselves.
