This week Microsoft made a significant internal announcement that did not get the same headlines as E7 or Copilot Cowork, but in some ways it is more revealing than either of them.
On March 17, Satya Nadella announced a restructure of Microsoft's entire Copilot operation. Consumer and commercial Copilot teams, which have operated separately, are being unified into a single organisation. Jacob Andreou, formerly of Snap, has been promoted to lead the combined effort. And Mustafa Suleyman, head of Microsoft AI, is stepping back from day-to-day Copilot work to focus exclusively on building Microsoft's next generation of AI models.
The announcement is worth paying attention to not because of the org chart, but because of what prompted it.
The Adoption Problem Microsoft Is Not Hiding
Microsoft 365 Copilot currently has around 15 million paying users. That sounds significant until you consider that the overall Microsoft 365 commercial user base is approximately 450 million. That means fewer than one in twenty commercial users with access to Microsoft 365 has a Copilot licence.
For a product that Microsoft has been marketing heavily for over two years, and which underpins a significant part of the company's AI revenue story, that number is low. Microsoft knows it. This restructure is a direct response to it.
The public framing from Nadella is forward-looking moving from a "collection of great products to a truly integrated system." But the underlying logic is simpler: Copilot has not yet become indispensable to most users, and Microsoft is reorganising to fix that.
What Is Actually Changing
The structural change is designed to remove friction between the consumer and commercial Copilot experiences. Until now, if you used Copilot on your personal phone it felt different to Copilot inside Microsoft 365. The roadmap, the design decisions, and the engineering teams were separate.
Unifying them means Microsoft is betting on a more coherent, consistent experience one where AI feels like a natural part of how you work across every surface, rather than a feature bolted on to individual apps.
At the same time, Suleyman's shift to focus on frontier model development is significant. Microsoft's current Copilot relies heavily on OpenAI models, and that dependency has been a topic of ongoing discussion in the industry. Building its own frontier models gives Microsoft more control over the direction, cost, and performance of its AI products over the next five years.
What This Means If You Are Evaluating Copilot
If your organisation has been sitting on the fence about Copilot, this restructure is worth factoring into that conversation in both directions.
On one hand, Microsoft is clearly doubling down. The investment in unifying the product, the E7 licensing bundle announced last week, the new agentic capabilities in Wave 3 these are not the moves of a company that is retreating from its AI position. The product is going to keep evolving, and the roadmap is ambitious.
On the other hand, the fact that adoption is this low after two years of availability tells you something important: deploying Copilot successfully is not simply a matter of buying licences. It requires readiness. Data needs to be in reasonable shape. Users need to understand what Copilot can actually do in their specific context. Governance needs to be considered before agents start taking action on behalf of your team.
The organisations that have gotten genuine value from Copilot are not the ones that bought licences fastest. They are the ones that took the time to understand where AI fits into how their people actually work, and built adoption from there.
The Bigger Picture
Microsoft's restructure, the E7 announcement, the July price increases, the arrival of agent capabilities across Power Platform and Microsoft 365 these are not isolated events. They are part of a consistent direction.
The Microsoft stack is moving toward a model where AI is not an add-on you purchase separately, but a core part of how the platform works. That shift has licensing implications, governance implications, and most importantly people and change management implications.
The organisations that navigate this well will not be the ones that react to each announcement individually. They will be the ones that have a clear view of where they are today, where they want to go, and what needs to be true before they can get there.
Microsoft also published its 2026 Release Wave 1 roadmap this week, outlining the next wave of capabilities across Power Platform, Dynamics 365, and Copilot Studio arriving between April and September. That deserves its own conversation more on that soon.
If you have been tracking these announcements and wondering what they mean for your Microsoft environment, that is exactly the conversation I am here to have.
Jan Davids Principal Consultant, Aureus Solutions Microsoft AI Cloud Partner | Adelaide, SA
Source: Microsoft Official Blog, March 17 2026 https://blogs.microsoft.com/blog/2026/03/17/announcing-copilot-leadership-update/
