Microsoft has reorganized its Copilot leadership in a move that signals the company is struggling to convert its massive AI investment into actual user adoption. Jacob Andreou, a former Snap executive, now leads the Copilot experience across consumer and commercial products, while Mustafa Suleyman, who joined Microsoft through its 2024 acquisition of Inflection AI, shifts focus to building enterprise-focused AI models. The restructuring reveals a fundamental problem: three years after launching Copilot commercially, only about 3% of Microsoft 365 business subscribers are paying for it. Why Is Microsoft Making This Leadership Change Now? The numbers tell the story. Copilot had approximately 6 million daily active users in February 2026, according to data cited by multiple sources. That pales in comparison to ChatGPT's 440 million daily users and Google Gemini's 82 million. Even Anthropic's Claude, a newer entrant, reached 9 million daily users by March 2026, nearly matching Copilot's user base. For a company that has invested billions in AI infrastructure and partnerships with OpenAI, these numbers represent a significant gap between ambition and reality. Investor patience is also wearing thin. As AI spending accelerates across the industry, stakeholders are demanding clearer paths to revenue and profitability. Microsoft's restructuring is partly a response to this pressure. By separating the work of building AI models from the work of making Copilot feel essential to users, the company is acknowledging that its previous organizational structure failed to translate technical capability into product momentum. What Does Jacob Andreou Bring to the Role? Andreou's appointment reflects a deliberate shift in priorities. Unlike traditional software products, AI assistants are judged on tone, consistency, speed, and whether they become habitual. Andreou's background at Snap, a company built on consumer engagement and growth, suggests Microsoft wants someone who understands how to make products feel indispensable rather than merely functional. He reports directly to CEO Satya Nadella, signaling the importance of the role. His mandate centers on adoption and retention. Copilot does not just need better underlying intelligence; it needs better cadence, better emotional resonance, and better framing in a market where "AI assistant" has become a commodity phrase. Andreou's consumer-first instincts are meant to address what Microsoft calls the "experience layer," the part users actually interact with every day. How Is Mustafa Suleyman's Role Changing? Suleyman, who co-founded DeepMind before joining Microsoft, is moving away from day-to-day Copilot product ownership. Instead, he will focus on what he calls "superintelligence efforts" and building "highly COGS-optimised, highly efficient enterprise-specific model lineages for Microsoft over the next three to five years". In other words, he is being asked to make Microsoft's AI models cheaper to run and more tailored to business use cases. This shift reflects a broader strategic tension within Microsoft. The company remains deeply dependent on OpenAI for cutting-edge models, with intellectual property rights secured through 2032. But Microsoft is simultaneously trying to build more of its own AI stack, including models, agents, and interface experiences that can stand independently. Suleyman's new focus on model development suggests Microsoft wants to reduce that dependency over time while improving the cost structure of its AI systems. What Are the Structural Problems Andreou and Suleyman Are Trying to Fix? Before the restructuring, Copilot's consumer and commercial products were developed by separate teams. This fragmentation created inconsistency and slowed decision-making. Suleyman acknowledged this directly in his internal memo, stating that combining these organizations "makes sense" because "every user, whether at home or at work, will be able to enjoy the full benefit of what we are all building". The new structure consolidates accountability while creating clearer lanes of responsibility. Alongside Andreou and Suleyman, Ryan Roslansky, Perry Clarke, and Charles Lamanna continue to lead Microsoft 365 applications and the Copilot platform, also reporting to Nadella. This arrangement is designed to reduce internal friction and speed strategic decisions. How to Understand Microsoft's New Copilot Strategy - Consumer Focus: Andreou's appointment signals that Microsoft is prioritizing user adoption and habit formation. The goal is to make Copilot a reflex, something people open without thinking, rather than a feature they occasionally use. - Model Layer Investment: Suleyman's shift to frontier model development indicates Microsoft is investing heavily in building proprietary AI systems that are cheaper to operate and tailored to enterprise needs, reducing long-term dependency on OpenAI. - Enterprise Governance: Charles Lamanna's oversight of the Power Platform and Copilot Studio reflects Microsoft's recognition that enterprise buyers care deeply about control, security, and customization. These tools allow organizations to build, customize, and govern how AI agents work inside their Microsoft 365 environment. - Agentic AI Expansion: The restructuring accompanies a product shift toward Copilot Tasks, Copilot Cowork, and Agent 365, tools designed to coordinate and execute tasks across Microsoft 365 rather than simply chat with users. The practical implications for organizations are significant. Copilot is moving from a tool that assists individuals to one designed to operate across teams and workflows. IT and workplace leaders who have not yet built governance frameworks, change management processes, and employee training programs for this shift are already behind. What Remains Uncertain? The restructuring does not guarantee that Copilot adoption will accelerate. Execution risk remains high. Microsoft must deliver consistent experiences across multiple product surfaces, including Windows, Microsoft 365, Bing, and developer tools. It must also navigate intense competitive pressure from Google, OpenAI, and Anthropic, each of which is investing heavily in AI assistants and agents. Enterprise buyers may also continue to resist premium pricing, especially if they perceive Copilot as a nice-to-have rather than a must-have. Model-building costs could rise faster than returns, and user confusion may persist if experiences remain inconsistent across surfaces. The dependency on OpenAI, while strategically important, also remains a potential vulnerability. What Microsoft is betting on is that separating the work of building models from the work of building products will unlock faster iteration, clearer accountability, and ultimately, better user experiences. Whether that bet pays off will depend on execution, competitive responses, and whether Copilot usage growth accelerates in the coming quarters. For now, the 3% adoption rate among Microsoft 365 subscribers suggests the company has significant work ahead.
