Microsoft is reshaping its engineering organization around agentic AI, bringing major developer and AI platform teams into a new division called CoreAI – Platform and Tools. The shift, announced by Microsoft CEO Satya Nadella, is designed to concentrate much of the company’s developer resources on the tools and infrastructure behind AI apps and agents.
A new center for Microsoft’s AI stack
CoreAI – Platform and Tools combines the existing AI platform team with the previous developer division, known for work that includes .NET and Visual Studio. It also brings in some key teams from the Office of the CTO.
In Nadella’s framing, the new group is not a narrow research unit or a side project. It is meant to build the end-to-end Copilot and AI stack for Microsoft’s own products and for outside customers building and running AI apps and agents.
The teams named in the reorganization include:
- Dev Div
- AI Platform
- AI Supercomputer
- AI Agentic Runtimes
- Engineering Thrive
The new division will also build out GitHub Copilot. That matters because GitHub Copilot is positioned as both a product and a source of feedback for the underlying AI platform. In other words, Microsoft wants the product experience and the platform roadmap to inform each other closely.
GitHub Copilot takes the spotlight
One notable signal in the announcement is what it emphasizes. The blog post does not focus on .NET or Visual Studio, even though the previous developer division was responsible for those tools. Instead, the attention is on GitHub Copilot, agentic AI and the broader stack needed to support AI applications.
That does not mean Microsoft is abandoning its established developer tools. The source article does not say that. But the structure of the announcement makes the company’s priority clear: the next wave of developer work is being organized around AI-first products and platforms.
For developers, the important point is that Microsoft is aligning tools, infrastructure and product feedback inside one group. The aim is to support both first-party and third-party customers as they build AI apps and agents.
Jay Parikh will lead CoreAI – Platform and Tools as EVP. Parikh was hired by Microsoft in October and previously worked as the VP and global head of engineering at Meta.
Why agentic AI is driving the reorganization
The source describes AI agents as applications with defined boundaries, or action spaces, and large memory capacity. Their purpose is to independently handle subsets of work that human office workers do today.
That idea sits between two competing expectations. Some company leaders and AI commentators believe these agents will replace jobs outright. Others take a more cautious view, seeing them as powerful tools that streamline work people already do.
Nadella’s own language points to a very ambitious view of the change. In his blog post, he wrote:
“2025 will be about model-forward applications that reshape all application categories.”
He also described the shift as affecting every layer of the application stack and compared the scope to several earlier platform shifts arriving together. His most direct summary was that “Thirty years of change is being compressed into three years!”
Those statements explain why Microsoft is reorganizing at this scale. If the company believes agentic AI will reshape software development and application categories, then keeping AI infrastructure, developer tools and Copilot work in separate lanes would make less strategic sense.
The broader bet behind CoreAI
Agentic AI is being presented as the next major turning point for the technology industry and for the world of work. The source contrasts it with general-purpose LLM chatbots like ChatGPT, which are not specialized in anything and require informed users to get the most out of them.
It is also described as different from limited-scope, feature-specific applications of deep learning, such as those Apple has introduced in its platforms and software. In this view, agents are not just chat interfaces or isolated features. They are a broader application model built around independent action within defined limits.
Most observers cited in the source expect some widespread use of agentic AI. The open question is how transformative it will be. Nadella’s reorganization suggests Microsoft is planning for the more aggressive scenario, where agentic AI becomes central to software development and to the industry around it.
The CoreAI structure is therefore both an internal management change and a statement of direction. Microsoft wants its software and services to underpin the AI apps and agents that customers build, while using GitHub Copilot as a close feedback channel for improving the stack.
The practical takeaway is simple: Microsoft is placing its developer organization inside its AI strategy, rather than treating AI as a layer added on top. For a company whose developer ecosystem includes long-running tools such as .NET and Visual Studio, that is a significant shift in emphasis.