Microsoft says daily Copilot users have grown nearly 50% in a quarter and more than fivefold year on year, while enterprise customers with over 50,000 seats have increased more than eightfold. Its earnings-call examples also show businesses using AI for operational work, from Unilever’s order processing to Rolls-Royce’s engine-maintenance planning.
Microsoft AI Watch analysis
What happened
Microsoft is pitching a move from AI experiments towards systems that handle business processes. At its fiscal first-quarter earnings call, CEO Satya Nadella pointed to Unilever, Rolls-Royce and Raiffeisen Bank International as examples. Raiffeisen, he said, has deployed AI across more than 20 business processes, including credit-related workflows.
Copilot use is growing alongside those deployments. Microsoft says automated tasks have nearly doubled since July, and customers with more than 50,000 Copilot seats are up more than eightfold year on year. The company is also adding consumption-based billing for agents, on top of per-user subscriptions. Customers using that billing grew by more than 75% in the quarter.
Why it matters
The useful change is not just more people trying Copilot. Microsoft is describing a shift towards AI handling repeatable business processes, while introducing a way to charge for agents doing more complex tasks. That could change both where companies deploy AI and how they budget for it.
The figures are Microsoft’s, reported by PYMNTS, and they do not show how much work the systems complete successfully or what those deployments cost. Microsoft CFO Amy Hood also expects usage-based pricing to spread more slowly outside software development, in part because it requires changes to sales practices and customer education.
Our read
This is a better adoption story than a parade of pilot projects: named businesses, specific tasks and substantial growth in Copilot use. But user numbers and billing growth are not the same as proven business value. The next test is whether these systems reliably deliver useful work at a cost companies will keep paying.
What to watch
- Whether Microsoft provides evidence of accuracy, time saved or business outcomes for these deployments.
- Whether consumption-based billing gains traction beyond software development.
- Whether the named companies expand these AI workflows or keep them to particular tasks.
Discussion spark: Should companies pay for AI agents by usage when the value of the work they complete is still difficult to measure?
Sources and evidence
- Source update (29 October 2026, 01:25 UTC)
- Source update (29 October 2026, 01:25 UTC)
not affiliated with or endorsed by Microsoft