MCP (Model Context Protocol)

Definition

Model Context Protocol: an open standard that lets an AI assistant connect to other systems — read their data, or carry out the actions they allow. The system the assistant connects to runs an “MCP server”; the assistant is the client. One assistant can connect to several servers at once, so a single conversation can draw on more than one source.

What it tells you

For a hotel, MCP is the difference between an AI assistant that answers from general knowledge and one that answers from your own figures. Connected to a revenue system, the assistant can look up last night’s pickup or this month’s position against budget instead of guessing. The connection does not make the answer correct by itself: the server supplies the data, and the assistant’s model still writes the answer.

How to track it

Not a metric, but a connection to check before you approve it. Can it only read, or can it also change things — rates, settings, emails? Which properties and data can it reach, and is that limited to your own access? Who in the hotel switched it on, and can it be revoked at once? And where does the requested data go: the assistant’s provider receives it, under that provider’s terms — on a company plan, conversations are generally kept out of model training.

Where it fits

The connection layer between AI assistants and hotel systems, next to the LLM that writes the answer and the AI agent that works through multi-step tasks. In Peaqplus, the Pulse AI Connector is a read-only MCP server: Claude or ChatGPT can query a hotel’s reports, forecast and budget tracking, within the user’s own access.