The Model Context Protocol (MCP) is an open standard that lets AI assistants such as Claude and ChatGPT connect to other software. An app that offers an MCP server exposes tools — search leads, create a task — that an assistant can call on your behalf.

Why would a travel agency connect an AI assistant to its CRM?

So the assistant works with real data: it can look up a client's enquiry, summarise open leads, check whether a quote was sent, or create a follow-up task — without someone copying information between the CRM and the chat.

Is it safe to give an AI assistant access to client data?

It depends on how access is controlled. In Tripdocks, assistants connect with OAuth 2.1, act as the user who connected them, are bound by that user's role, have every call logged, and can be revoked instantly.

Which assistants work with Tripdocks?

Claude (including Claude Desktop), ChatGPT on plans that support custom MCP connectors, and any client that supports remote MCP servers over OAuth.

Learn · AI

MCP for travel agencies

Your AI assistant is only as useful as the data it can reach. MCP is how it reaches your CRM — safely.

By the Tripdocks team · Last updated September 28, 2026

MCP — the Model Context Protocol — is an open standard for connecting AI assistants to other software. For a travel agency, it means an assistant like Claude or ChatGPT can work with your real leads, quotes, and tasks instead of whatever you paste into the chat.

What an assistant can do once it's connected

"Which leads for Bali came in this week and haven't had a reply?"
"Summarise everything we've sent the Sharma family."
"Create a follow-up task for tomorrow on the Kerala enquiry."
"Log the advance payment against this reservation." (only if your role allows it)

In Tripdocks, a connected assistant can read leads, contacts, pipeline, quotes, itineraries, reservations, tasks, and invoice and payment status — and, if you grant it, create or update leads, manage tasks, send quotes, and log payments.

What to check before connecting any AI to client data

Question	Why it matters	How Tripdocks does it
Does the assistant get a password?	Shared passwords can't be scoped or revoked	OAuth 2.1 with PKCE — scoped tokens only
Can it see more than the user?	Assistants shouldn't widen access	Acts as you, bound by your role and data scope
Is there a record of what it did?	You need an audit trail	Every tool call logged; arguments hashed
Can you switch it off?	Access must be revocable	One click in Settings → Intelligence
Can admins see all connections?	Owners need oversight	Workspace view for admins with Manage AI access

How to connect

Make sure your role has the Connect AI permission.
In your assistant, add a custom connector pointing at https://app.tripdocks.com/mcp.
Sign in to Tripdocks on the consent screen, review the access, and allow it.

Step-by-step guides: Claude, ChatGPT, and the AI assistants documentation. For the bigger picture, read why travel software has to become AI-native.

Frequently asked questions

What is MCP?: The Model Context Protocol (MCP) is an open standard that lets AI assistants such as Claude and ChatGPT connect to other software. An app that offers an MCP server exposes tools — search leads, create a task — that an assistant can call on your behalf.
Why would a travel agency connect an AI assistant to its CRM?: So the assistant works with real data: it can look up a client's enquiry, summarise open leads, check whether a quote was sent, or create a follow-up task — without someone copying information between the CRM and the chat.
Is it safe to give an AI assistant access to client data?: It depends on how access is controlled. In Tripdocks, assistants connect with OAuth 2.1, act as the user who connected them, are bound by that user's role, have every call logged, and can be revoked instantly.
Which assistants work with Tripdocks?: Claude (including Claude Desktop), ChatGPT on plans that support custom MCP connectors, and any client that supports remote MCP servers over OAuth.