You got the Dynamics 365 project. The requirements are clear. More or less. In modern enterprise solutions, it is quite common to have a rich data model with a lot of different tables entwined with each other in various ways. Which naturally creates lookups in forms. This creates connections between records giving the opportunity for a 360-degree view of your customers, leads, opportunities, projects etc. Or in my geeky case of the Avenge 365 app based on Dynamics 365 Sales,

In the last blog post, I covered how Copilot Studio really accesses Dataverse through various lenses with the Dataverse Knowledge Source, Dataverse Connector and Dataverse MCP Server. From the way the core model works, to the way inputs and outputs are managed, schema awareness, CRUD support and bound/ unbound actions. Here is what the summary looked like: Capability Knowledge Source Connector MCP Server Core Model Search/ index based, for retrieval + grounding Deterministic

There’s a moment in every agentic strategy project where you realize: how your agent connects to data matters just as much as what data it sees. Now, coming from the world of Dynamics 365 and Model-Driven Power Apps, Dataverse is the golden data source. So naturally, when there is an ask to create a custom agent which can “speak” the business app’s language, what is the next step? Connect to Dataverse of course! Looking at your metaphorical whiteboard, three (plus a bonus one