AI assistants (MCP)
Letting your team and your clients connect Claude or another AI assistant to their data, and governing what it can reach.
-
AI assistants and the MCP server
What the MCP server is, who it is for, and the one rule that answers most questions about it.
-
Turning MCP on
The four things that must all be true before an assistant returns any data, in the order to check them.
-
Connecting an assistant
The server address, the setup steps for Claude, Claude Code, ChatGPT and Gemini, and what to expect from the consent screen.
-
What an assistant can do
The tool catalog, so you can answer can it do X precisely, plus how filtering, sorting, and paging work.
-
Changing data safely
Every write is a preview first, which fields are refused by design, and how bandwidth change orders work over MCP.
-
Governing connected assistants
Reviewing and revoking a connected assistant, capping one at read-only, and which levers are yours versus Modero's.
-
Getting good answers
The habits that separate a useful MCP session from a frustrating one, and prompt patterns worth passing on.
-
When MCP is not working
Causes ordered by how often they turn out to be the real one, the edge-block diagnostic, and what to capture before escalating.