Microsoft now charges $0.01 per Copilot Credit for agents such as Cowork, and that single line item is quietly rewriting the economics of Microsoft 365 Copilot for enterprises. Three years ago the pitch was simple: a flat $30 per user per month, with a $21 tier for smaller businesses added later. The pitch has since acquired a meter.

Usage-based pricing questions now rank among the most common Copilot queries Gartner fields from clients. Max Goss, research director at Gartner, detects "some doubt about whether a full move to consumption-based pricing for everyday knowledge work is sustainable for organizations in the long run." He is right to hedge. Doubt is the rational position.

What did Microsoft actually change?

At a recent event the company unveiled a new Home interface that places Cowork alongside the standard Chat option in the prompt box. Chat runs on the subscription. Cowork burns credits, with consumption varying by model and task complexity, according to Microsoft's pricing documentation.

The distinction sits one click apart. Goss warns that "bringing these together in a single interface is going to make this harder" — an employee could summarise calendars in Cowork, "potentially burning tokens on a task Copilot can do for free." Anthropic has taken a similar path, merging Claude's chat and Cowork tools, though Claude Enterprise meters both.

Microsoft says it is building automatic routing to decide whether a query goes to Cowork or Chat. Keith Kirkpatrick, research director at The Futurum Group, sees the conflict of interest: "When vendors decide whether a task should fall into one bucket versus another, there's a risk that they're overly weighted to moving the task to the bucket where they're going to monetize it." Quite.

Is consumption pricing survivable at 10,000 seats?

GitHub Copilot developers already pay by consumption, and that works — narrowly. "Consumption-based pricing for a specific task like development is plausible and sustainable, albeit with some caveats," said Goss. "The big question — that is still an open question — is: is consumption-based pricing scalable for general purpose knowledge work?"

Multiply an agent invocation in a spreadsheet by 10,000 or 20,000 employees, factor in repeated retries from less proficient users, and the invoice stops being a forecast. A study cited in the Wall Street Journal found just 11% of surveyed firms successfully predicted spending on agentic systems.

Microsoft's cost management tool lets admins set organisation-wide and individual spending limits, monitor credit balances and trigger notifications at thresholds. Useful, but monitoring is not forecasting. "The challenge is around forecasting," said Kirkpatrick, since adopters lack data on how employees will actually use the tools.

What does this change for IT and finance teams?

The practical answer is FinOps, whether anyone asked for it or not. "Successful organizations are the ones that have devolved this," said Goss, with IT retaining oversight while business units appoint credit "stewards" operating within a fixed pool. He is blunt about readiness: "IT procurement and finance teams will have to work closer together in more of a technical FinOps model. Most companies aren't there yet."

There is also the value question, which never really went away. A Gartner survey of 186 IT leaders published in June found 51% cited value as a top-three barrier to Copilot deployment. Respondents reported high value from content generation (49%) and knowledge search (44%), but only 20% said the same of automating workflows with agentic tools such as Copilot Studio.

Hand employees Cowork access with a "go for it," and Goss predicts a relapse: "they're almost certainly going to see wastage. It's going to be back to the early days of Copilot, when people were asking: 'Is this valuable?'"

Kirkpatrick's advice is to define business outcomes first: "If you just focus on the technology, you're going to wind up trying things out … without any real understanding of whether or not there's actually business value being delivered." He recommends pilots before broad rollout. Goss suggests small Cowork trials with fifty experienced users, strict credit caps and clear guidance on when Cowork beats plain Chat.

One negotiation lever remains. The new Microsoft 365 E7 tier bundles M365 with Copilot and Agent 365 but excludes Copilot Credits — deliberately, one assumes. Goss suggests tying the two conversations together: "If you are going to move to E7, what can you get? Microsoft could technically assign you credits, even if it's just a starter amount."

Verdict: pilot Cowork with limits, brief employees on Chat-versus-Cowork, and renegotiate before the meter starts. The $30 subscription was predictable. The credit balance is not.