Have any large organizations stopped using Microsoft CoWork since the July 1 charge?

A restricted rollout makes more sense than giving everyone access by default. One large pilot used roughly $90,000 of a $100,000 credit across 400 to 500 users, with the top 10% accounting for most of the usage.

That kind of concentration makes a simple per-user average misleading. Start with specific tasks, track what they cost, and compare that with the time or money saved before expanding access.

Keep it opt-in and give access to people with a concrete reason to use it. A five-person team is a very different cost decision from an organization-wide rollout, and there's not much point enabling it for people who won't use it.

Cost isn't the only reason to pause it. Security and governance matter too, especially if the tool can carry out tasks across an enterprise environment. Figure out what it should be allowed to do and who is responsible for reviewing those uses before opening it up broadly.

I'd also separate general chat from work that actually needs CoWork. Some people may reach for it because they prefer its responses, even when Copilot Chat is adequate. Reserving CoWork for tasks or agents with a clear benefit could help control spending.

I'd be careful about projecting a full rollout from a small pilot. One reported test cost about $24 for three prompts, but prompt count alone doesn't explain the complexity of the work or establish a typical price. Measure actual workloads before committing to an annual budget.