The Data Quality App Now Speaks MCP: Clean Data for the Agent Era

Today we are announcing support for the Model Context Protocol (MCP) in the Data Quality App. AI agents, from Microsoft Copilot Studio to Claude, can now talk to your data quality layer directly: ask about the health of your Dynamics 365 data, check records for duplicates, and put clean data at the centre of every automated decision.

This post covers three things: what MCP actually is, what it does for the Data Quality App today, and why it matters for where Microsoft is taking Dynamics 365, a future where agents, not forms, do the work.

What is MCP?

The Model Context Protocol is an open standard that defines how AI assistants and agents connect to tools and data. Think of it as USB-C for AI: before USB-C, every device needed its own cable. Before MCP, every AI integration needed its own custom connector, built and maintained separately. MCP replaces that with one plug that works everywhere.

The protocol has two roles. An MCP client is the agent, Copilot Studio, Claude, ChatGPT, or any AI assistant that speaks the protocol. An MCP server is the tool side: an application that describes its capabilities in a standard way, so any agent can discover what it offers and use it in a controlled manner.

MCP was introduced in late 2024 and has since been adopted across the industry, including by Microsoft, which supports MCP in Copilot Studio and ships its own MCP server for Dataverse. When Microsoft, OpenAI, Google and Anthropic all standardise on the same protocol, it stops being a trend and becomes plumbing: the way agents and business systems will talk to each other for the next decade.

What MCP means for the Data Quality App today

The Data Quality App now ships with MCP support. Connect an MCP-capable agent to your environment, and it can work with your data quality layer the same way a colleague would:

• Ask about data health. Open duplicate pairs, records merged, rule compliance, data health per table over 30, 60 and 90 days, the numbers from Data Health Studio, available to any agent that asks.
• Check records for duplicates. Exact, fuzzy and phonetic matching on demand, so an agent can check before it creates or updates a record, not after the damage is done.
• Run and review duplicate detection. Trigger detection on a table and work with the results as part of an automated flow.
• Consult your data quality rules. Which requirements apply to a record, and whether it meets them, the same rules your team defined, now readable by agents.

In practice it looks like a conversation: “How healthy is our account table this month?”, “Data health is at 84%, up 3 points. 412 open duplicate pairs, most from last week’s trade-show import. Want me to run detection and prepare a merge review?”

And because the Data Quality App runs natively inside your own Dynamics 365 environment, nothing changes about where your data lives. You decide which agents connect, and MCP gives them a defined, inspectable set of capabilities, not open-ended access.

Why this matters: Dynamics 365 is becoming an agent platform

Microsoft has been explicit about the direction. Copilot is now the front door to Dynamics 365, Copilot Studio lets every organisation build its own agents, and Microsoft’s leadership talks openly about a future where business applications are operated by agents rather than navigated by people. Sales agents that qualify and follow up on leads. Service agents that resolve cases end to end. Operations agents that keep master data moving between systems.

In that world, the forms and grids fade into the background. What remains is the data, and software that acts on it autonomously, at machine speed, around the clock.

In an agent-run CRM, data quality is safety

Agents multiply whatever data you give them. A human seller who spots two records for the same customer hesitates, checks, asks a colleague. An agent does not hesitate. It emails both records. It logs activity on the wrong one. It reports pipeline twice. Every data quality problem you have today gets executed faster and more consistently once agents take over the keyboard.

That flips data quality from housekeeping to infrastructure. Duplicate prevention, data requirements on every entity, and measurable data health stop being a quarterly cleanup project, they become the guardrails that decide whether your agents are an asset or a liability.

This is exactly why we built MCP support. It puts the Data Quality App in the loop before agents act: check the record first, respect the rules, merge instead of duplicate, and keep a live score of whether the foundation is getting better or worse.

Where we are taking this

This first release is the foundation. On our roadmap: agents that monitor Data Health Studio and raise cleanup work on their own, prevention that gates what agents are allowed to write, not just what users type, and deeper scenarios for Copilot Studio builders who want data quality as a ready-made building block in their agents.

Get started

MCP support is available for the Data Quality App now. If you are piloting Copilot or building agents on Dynamics 365 data, start where the agents will start: with the data itself. Book a demo and we will show you MCP against a live environment, or read what a complete data quality solution includes first.