Glossary
Definition
The Model Context Protocol is an open standard that lets AI models and agents discover and call external tools and data sources through a common interface.
The Model Context Protocol defines how an AI client (a model or an agent) connects to a server that exposes tools, resources, and prompts. A server describes what it offers with schemas; the client discovers those offerings and calls them. Because the interface is standard, one tool server can serve many agents, and one agent can use tools from many servers, without custom integration for each pair.
For enterprises, MCP is the practical path to letting agents from different vendors reach the same systems under one policy. The tool server is where scopes, validation, and rate limits live, so governance is applied where the call happens rather than inside each agent.
Meridian's Gateway exposes HR and finance systems as MCP servers with per-agent scopes and approval holds, and accepts any MCP client that authenticates through the customer's identity provider.
Read next
Long-form explainers and frameworks that use this term, dated and signed.
Explainers
MCP explained without jargon: how agents discover and call tools, where tool permissions must be enforced, why the gateway pattern matters, and how OpenTelemetry makes agent runs observable.
Owen Gallagher6 min read
Governance
A seven-part framework for governing AI agents in HR: scope statements, least-privilege data permissions, approval tiers, audit trail, model choice, employee communication, and works council and GDPR alignment.
Marcus Feld7 min read
Get started
Start with one workflow, one approver, and one number to move. Most design partners were live in five weeks.