Up until now, Microsoft’s AI solution, Copilot, has been useful – but somewhat limited when used with Dataverse (the underlying data layer for Dynamics 365 and Power Apps).
It could see emails, meetings, documents and chat threads, but it couldn’t see the systems where operational data lived (Dataverse). It had no knowledge of accounts, opportunities, case histories, or the information buried inside the apps businesses rely on every day.
With Microsoft’s announcement that Copilot can finally work directly with Dataverse, that gap has now closed.
For organisations using Dynamics 365 or custom Power Apps, this is the moment where AI becomes genuinely practical because instead of guessing or relying on what’s visible in Outlook and Teams, Copilot can now draw context from your wider business applications and vice versa.
Gavin Nixon, Managing Director of Strategy 365, said: “Copilot originally used Microsoft Graph as one of its key data sources. Then Microsoft introduced Work IQ to give Copilot deeper organisational understanding of that Graph data. Now, with Dataverse integration, Copilot can also access extended business data from Dynamics 365 or custom Power Apps, providing a far broader view of your organisation, all working together for the first time.
“So, if an AI agent is only as smart as the data it can reach, then agents just got a whole lot smarter! This really marks a huge turning point for Copilot, and we’re excited to see where it will take us.”
A different day to day experience
Instead of moving between Dynamics 365, Power Apps, Outlook and Teams, the user stays inside the main Microsoft 365 Copilot app. A sales opportunity can be created directly from a meeting summary using the connection to Dynamics 365 Sales. An account overview can be generated without searching through dashboards or reports. A question about pricing changes can be answered with the correct opportunities surfaced instantly. The user doesn’t need to know which screen the answer is on. They simply ask, and Copilot brings together the operational record and the wider context.
This also applies to custom-built apps (which we’re a huge fan of here at Strategy 365 due to their cost-effective licences). An HR team can call on their HR app, built with Power Apps, right within Copilot and ask it to consolidate hiring figures. Another example could be a service representative starting their day by reviewing priority cases without opening Dynamics 365 Customer Service. Actions happen in one place, but they’re grounded in the organisation’s actual data, not a simplified version of it.
Gavin added: “For now, Copilot, Work IQ and Dataverse together represent the closest thing we’ve seen to an AI that genuinely understands an organisation’s data in a meaningful way. That combination is what makes this change stand out among the developments announced over the past year.”
Where this becomes especially significant is in roles that rely on accurate record keeping. A sales representative, for example, no longer needs to manually summarise a customer call. If the call is held in Teams, the transcript provides a complete account of what was discussed. Combined with emails, documents and previous activities, Copilot can surface a level of detail that often gets missed simply because people are busy – then it can update the record in Dynamics 365. It reduces the risk of incomplete updates and creates a more accurate, unified view of each relationship.
How does it work?
The element that makes this possible is Work IQ. Microsoft describes it as the layer that resolves entities and relationships across structured and unstructured data. In practice, it connects the signals from Microsoft 365 (the things people talk about in meetings, emails and documents) with Dataverse records such as accounts, leads, activities and any custom entities a business has created. If a pricing change is discussed in a meeting, Work IQ knows how that decision relates to live opportunities in Dynamics 365. It understands which records are relevant, and what needs action.
The simple explanation is that Copilot can now combine two previously separate worlds: the operational records stored in Dataverse, and the ongoing conversations stored across Microsoft 365. This is the first time Copilot has had both sides available at once.
Will Copilot interpret intent more reliably?
Microsoft has said the reasoning still involves schema and keywords. That point is worth acknowledging, because anyone who has ever struggled to find something in Outlook knows how dependent search can be on the exact words used. The real test will be whether the additional Dataverse context helps Copilot interpret intent more reliably. Strategy 365 will be watching that closely, particularly in cases where customers use custom tables or non-standard field names. At the moment, it’s not known whether Copilot will fully understand those structures or whether certain naming conventions will matter.
What it means for developers
For organisations building custom business applications, this will also change how solutions are designed. Microsoft is already pointing towards agent driven apps built with Copilot Studio, GitHub Copilot, or Azure AI Foundry, all able to work directly with Dataverse via the MCP server. Instead of static data entry screens, the focus shifts to assistants that understand the data model, the user’s responsibilities, and the context of the work they are doing.
When can we give it a go?
Public preview is rolling out across Power Apps, Dynamics 365 Sales and Dynamics 365 Customer Service from early April 2026, with additional integrations coming through Microsoft’s partner ecosystem. A Microsoft 365 Copilot licence will be required, and some scenarios involving Power Apps will also need a Power Apps Premium licence.
If the capability develops in the way Microsoft is positioning it, many of the manual steps people currently take, like digging through emails, updating records, and piecing together what happened last week, will start to fall away.