Using 'trillions of anonymised productivity signals', Microsoft thinks nearly half of people like AI in their work
Recently, Microsoft put out a new blog that attempts to educate/sell the notion of AI in the workplace, and perhaps most strangely, it admits to using huge swathes of anonymised data in the process. The blog, titled "How Frontier firms are rebuilding the operating model for the age of AI", says that authors, editors, directors, and orchestrators will all see the benefits of AI once they start using it.
It's the "what the data shows" section that caught my eye, though. It declares Microsoft analysed "trillions of anonymised Microsoft 365 productivity signals and surveyed 20,000 workers using AI across 10 countries."
The main bulk of Microsoft's report comes from a 'privacy-preserving analysis' of over 100,000 chats in Microsoft 365 Copilot, and it found 49% of conversations support "cognitive work", and 58% of AI users say they now create work they couldn't have a year ago.
As part of this research, Microsoft found that the majority of AI users want to use AI for quality control and critical thinking. I can understand the rationale for the former, but replacing critical thinking with AI feels like a good way of reducing your critical thinking capacity in general.
Just last year, a Microsoft co-authored paper suggested the continued and regular use of generative AI left users with a "diminished skill for independent problem solving", which certainly spells bad things for those looking to replace critical thinking.
One element that is relevant to AI in the workplace is FOMO. Microsoft's study suggests that 65% of AI users fear they will be left behind if they don't adapt to AI quickly.
Xbox needs to move faster, deepen our connection with the community, and address friction for both players and developers.Today, we promoted leaders who helped build Xbox, while also bringing in new voices to help push us forward. This balance is important as we get the business…May 5, 2026
No word is given on exactly how this 'privacy-preserving analysis' is done, but I'm certainly feeling a tad cautious talking about AI in Microsoft products going forward—especially given recent discoveries about how other Microsoft software handles data. One would assume Microsoft 365 Copilot means Copilot, but it's sort of the umbrella term for what used to be Microsoft 365, so we don't know which chats it analysed.
Frontier firms, as talked about by Microsoft, are AI organisations led by humans, using new agentic AI. It's good for Microsoft and many other AI providers if these companies come about because the humans in them need those tools to function. And likely, they would pay a high premium for that privilege.
AI is absolutely everywhere, and Microsoft is one of many companies looking to cash in on it. Nvidia, one of the biggest providers of AI hardware and one of the biggest supporters of AI software, became worth $5 trillion last year. The software giant has been critiqued for putting AI in its software for some time, but Asha Sharma, the new Xbox CEO, has recently announced the winding down of Copilot on mobile and the cancellation of Copilot on console. And the company has been rolling back at least some AI features on Windows.
Microsoft is clearly still in on AI and wants you to be too, but it is slowly shifting how it talks about the tech and the productivity gains it wants to sell you on as a result.