New to MCP Server? Here are answers to the most common questions about AI-powered territory planning through Model Context Protocol.
What is an MCP server for territory planning?
The Model Context Protocol (MCP) is an open standard that lets AI agents connect to external tools and data. MCP Server exposes 47 territory planning tools — building, balancing, routing, scheduling, analysis — to any MCP-compatible AI agent like Claude Desktop or OpenClaw. Instead of clicking through a UI, you describe what you want in natural conversation and the agent executes it through the MCP Server.
Which AI agents work with EasyTerritory MCP?
Any compatible host works, including Claude Desktop, Claude Cowork, OpenClaw, and Cursor. Hosts that can set a header use an API key. Hosts that only support OAuth sign in at the server and paste the same key once on the consent page. See Connecting Your AI Agent and Connecting an OAuth-Only Agent.
See the Connecting Your AI Agent guide for step-by-step setup instructions.
Is my data stored by the MCP server?
No. MCP Server never retains your customer data. Account data is ingested, processed for the requested operation, and returned as standard GeoJSON. You and your agent own the data end-to-end. This eliminates the risk of a third-party becoming the system of record for your critical internal territory information.
Do I need to install or self-host anything?
No. EasyTerritory hosts the MCP Server — you connect your AI agent to the endpoint and go. There is no infrastructure to deploy, no dependencies to install, and no server to maintain. If you prefer hands-on refinement, export the Territory Solution and open it in Territory Designer.
Can I use my existing territory files?
Yes. Account lists come in as CSV (or inline rows). The durable Territory Solution is a standard GeoJSON FeatureCollection — save with export_geojson, reopen with import_geojson. There is no compact or proprietary project format. If you have an existing alignment in a spreadsheet (Postal Code + Territory columns), the direct_build tool maps it automatically.
How does AI territory balancing work?
You define the goals — territory count, target workload hours, or balance metrics — and the agent uses spatial optimization to partition accounts into balanced territories. The server considers drive time, account density, and workload distribution. After building, you can analyze equity across territories, then split, merge, or realign individual regions through conversation or visual map selection.
See Territory Building with AI for detailed examples.
What is the difference between Territory Designer and the MCP Server?
Territory Designer is a browser-based visual application for building and managing territories with a point-and-click interface. The MCP Server exposes the same territory planning capabilities through the Model Context Protocol, letting AI agents do the work through natural conversation. Both share the same underlying engine. Save the Territory Solution as GeoJSON and open it in Territory Designer for hands-on refinement. A Designer project can be brought back through GeoJSON import.
What data formats does the MCP Server accept?
CSV for account ingest. GeoJSON FeatureCollection for the Territory Solution (the only project file). Gzip is compression, not a second format. Account data can include any columns (revenue, units, rep name, etc.) — these become point properties available for balancing, classification, and analysis. The server geocodes addresses inline during ingest if coordinates are missing.
How many tools does the MCP Server expose?
47 tools, organized into two tiers. Tier 1 goal tools (prefer these) handle complete workflows in one call: build, ingest, realign, analyze, route. Tier 2 primitives compose for edge cases: part selection, map configuration, GeoJSON import/export, delegation. See the full Tools Reference for the complete list.
Can I build territories for a specific state or region only?
Yes. When building, set part_scope to "explicit" and provide a part_filter (e.g. {"state_abbr": "TX"}). This ensures the build only uses parts within the specified state, rather than inferring scope from the bounding box of your account data.
What if my agent only supports OAuth?
Add https://mcp.easyterritory.ai as a custom connector and sign in. On the consent page, paste the API key EasyTerritory issued. You enter it once. Leave any OAuth client ID and secret blank. The walkthrough is Connecting an OAuth-Only Agent. No key yet? Request a free trial.
Can I grow one territory around a single location?
Yes. seed_build grows one territory outward from an address or coordinate until a location count or a metric total is as close as possible. It does not divide every account into balanced territories, and it does not draw a drive-time ring. Ask for one territory per location.
What is the ExpertPack knowledge layer?
The ExpertPack is a structured knowledge layer that encodes a decade of EasyTerritory’s territory design expertise — workflows, tool selection rules, common mistakes, and design decisions. It is not found in foundation AI models. The MCP Server uses it to guide agents toward correct workflows, reducing wrong-tool and wrong-order failures.
Still have questions? Start a free trial or contact us — we are happy to help you get started.
