Microsoft's new AI tracking tool lets bosses see who's (not) using Copilot

Organisations will be able to compile AI engagement leaderboards

Redmond doubles down on workplace AI adoption with new Viva Insights benchmarks, but not everyone's convinced it will help productivity.

Microsoft is giving managers a new way to measure how enthusiastically their teams are embracing AI – and who's lagging.

The tech giant has updated its Viva Insights platform to include Copilot adoption benchmarks, allowing companies to compare AI usage across teams and even against other businesses.

"We're excited to announce the initial rollout of Benchmarks in the Microsoft Copilot Dashboard in Viva Insights," the company said.

"This new feature lets organizations compare Copilot usage internally across different company groups, as well as externally against similar companies."

Microsoft Copilot is Microsoft's AI assistant — a suite of AI-powered tools integrated across the company’s products to help users work, write, code, and communicate more efficiently. It's essentially Microsoft's version of a personal AI coworker, designed to sit inside familiar apps like Word, Excel, Outlook, PowerPoint, Teams, and Windows.

According to the company, the new Benchmarks feature, part of Microsoft's broader Viva suite for employee experience and analytics, will show managers detailed metrics such as the percentage of Copilot users, adoption rates by app, and returning user percentages.

Organisations will also be able to compare adoption based on variables like manager type, region, or job function – effectively creating a leaderboard of AI engagement across departments.

Externally, companies can now see how their AI adoption rates stack up against industry peers, something Microsoft says will "provide broader context and new opportunities to improve Copilot engagement."

External benchmarks are generated using randomised statistical models to protect privacy. Each benchmark group represents data from at least 20 companies and relies on aggregated estimates, ensuring that no individual company's actual information is disclosed.

The message from Redmond is clear: we want to you to make the use of Copilot a key performance indicator.

Microsoft has long marketed Copilot as an indispensable productivity booster, but more recent research suggests those claims may not hold up under scrutiny.

In July 2025, a study by the nonprofit Model Evaluation & Threat Research (METR) found that AI coding assistants actually slowed down experienced developers.

The research observed 16 veteran open-source programmers completing 246 real-world tasks using AI tools.

Before starting, participants estimated that AI would make them 24% faster. Afterward, they still believed their productivity had improved by about 20%. But the data showed the opposite: developers took 19% longer to complete their tasks when using AI.

The findings ran contrary not only to the developers' perceptions but also to predictions from economists and machine learning experts, suggesting that AI's perceived productivity boost may be an illusion - at least for seasoned professionals.

Similarly, an October 2024 study found that workplace AI tools failed to reduce burnout or meaningfully improve productivity, while some workers reported feeling increasingly like "assembly line operators" as AI systems optimized for speed and output rather than creativity and collaboration.

A joint Stanford University Social Media Lab and BetterUp Labs study released in September 2025 found that 40% of US employees reported encountering what they called "workslop" – low-quality, AI-generated material cluttering their workflows.

According to the study, sorting through such AI "slop" costs employers an average of $186 per employee per month in lost productivity.

The research also highlighted a troubling decline in workplace trust, as employees struggle to distinguish between human and machine-generated content.