Microsoft's aggressive multi-billion-dollar push into artificial intelligence has produced a sobering reality: less than 3% of paying users actively use Copilot, despite the company embedding it directly into Windows 11 and Microsoft 365. According to Mat Velloso, a former Vice President at Microsoft who spent 12 years at the company and four years as Technical Advisor to CEO Satya Nadella, the company has fundamentally failed to connect AI features with real users, prompting calls for a complete internal overhaul.

What's Actually Happening With Microsoft's Copilot Adoption?

The numbers tell a stark story. Out of Microsoft's 450 million Microsoft 365 users, the company has converted only roughly 15 million to paid Copilot seats. This translates to a 3.3% paid adoption rate, meaning 96.7% of users are rejecting the premium AI features. When measured against Microsoft's estimated $37.5 billion quarterly AI spending, this adoption rate represents an alarming disconnect between investment and user demand.

Velloso, who later moved to leadership roles at Google DeepMind and Meta's Superintelligence Labs, observed that Microsoft's behavior over recent months has shifted dramatically. Both the Windows and Xbox divisions suddenly began prioritizing user feedback and implementing requested features after years of ignoring them. This pivot suggests the company is finally confronting the reality that its AI strategy isn't resonating with customers.

The problem extends beyond software. Microsoft has heavily pushed original equipment manufacturers (OEMs) to include Neural Processing Units (NPUs) in their latest laptops to power advanced Windows 11 capabilities. However, Velloso noted that OEMs invested heavily in NPUs only to discover that "nobody cares because not a single valuable usecase was built for those in Windows/Office". This represents a significant misalignment between hardware investment and actual user needs.

Velloso

Why Is Microsoft Struggling When It Has Such a Large User Base?

The company's flagship AI search product, Bing, provides another cautionary tale. Despite making Bing its biggest AI bet, Microsoft failed to capture even a single percentage point of search market share from Google. This failure suggests that simply integrating AI features into existing products is insufficient when users don't perceive clear value.

Additional infrastructure challenges compound the problem. GitHub, which should be thriving as a centerpiece of the AI coding revolution, has seen its Service Level Agreement (SLA) reliability drop below 90%, while rising costs of goods sold (COGS) are putting pressure on margins. These operational issues undermine confidence in Microsoft's ability to deliver reliable AI services at scale.

The strain appears to be taking a toll on Microsoft's leadership structure. Julia Liuson, the highly respected head of Microsoft's Developer Division after 34 years with the company, recently retired. While official channels framed this as a standard retirement, Velloso critiqued the news, saying Microsoft has "just went from hit refresh to hit factory reset". He also documented a string of high-profile departures and reassignments across Xbox, GitHub, AI Infrastructure, Teams, and OneNote divisions.

How Should Microsoft Address Its AI Strategy Problems?

• Conduct a Complete Internal Reset: Velloso argued that hitting a complete internal "factory reset" is the only pragmatic way to fix the hard problems plaguing the operating system and enterprise software stack, rather than continuing with incremental changes.
• Stop Forcing AI Features on Users: Instead of pre-deploying Copilot into the Windows 11 taskbar and Office suite, Microsoft should focus on building genuinely valuable use cases that users actively want to adopt.
• Rebuild Enterprise Consulting Capabilities: Microsoft should invest in the hands-on, enterprise-level consulting that historically built its penetration into large companies, as competitors like OpenAI are now replicating this playbook with their own Forward Deployed Engineers.
• Address Reliability and Cost Issues: The company needs to improve GitHub's service reliability and manage rising costs of goods sold to maintain competitive margins in the AI services market.

The competitive landscape is shifting rapidly. OpenAI recently launched the "OpenAI Deployment Company" (DeployCo), a new business unit backed by over $4 billion in initial investment, featuring 150 "Forward Deployed Engineers" tasked with embedding directly into Fortune 500 companies. This move threatens Microsoft's historic dominance in enterprise software by cutting out the middleman and owning the direct relationship with businesses. As Velloso noted, Microsoft's incredible penetration into large enterprises was historically built on "armies of people spending time, listening, understanding business goals and solving them with technology in every industry vertical".

As Velloso

"Microsoft missed the internet wave, the mobile wave and now it missed the AI wave. It is what happens when you keep doing the same things expecting different results," stated Mat Velloso.

Mat Velloso, Former Vice President at Microsoft

When Frank X. Shaw, Microsoft's Lead Communications executive, defended the departing executives and accused Velloso of applying a "negative frame" onto normal corporate retirements, Velloso pointed to harsh financial realities. He noted that since moving from Microsoft to Google in early 2024, Google's shares surged by roughly 230%, while Microsoft's stock growth remained essentially flat at 0%.

Is Microsoft Actually in Danger?

Despite these severe criticisms, Velloso defended his former employer against apocalyptic narratives suggesting the company is dying. He acknowledged that while AI startups and labs are building flashy deployment companies, completely replacing legacy enterprise software is incredibly difficult. "There's a reason why all the top AI labs are hiring large consulting teams," he explained. "The last mile is the hardest and Microsoft has the best distribution for that. Their moat is unbreakable".

This creates a paradox for Microsoft in 2026. The company simultaneously faces severe internal challenges with AI adoption and user engagement, yet retains structural advantages in enterprise distribution that competitors cannot easily replicate. The question is whether Microsoft can leverage these advantages to rebuild its AI strategy before competitors like OpenAI solidify their direct relationships with Fortune 500 companies. The coming months will reveal whether the company's leadership can execute the kind of fundamental reset that Velloso and others are calling for, or whether Microsoft will continue down a path of incremental changes that fail to address the core problem: users simply don't want the AI features Microsoft is forcing upon them.