OpenAI finally hands the enterprise a spend dashboard - and budget caps

OpenAI this week gave administrators a way to see and cap what their organizations spend on AI. A refreshed admin console now pulls consumption for both ChatGPT and its Codex coding tool into a single view, sliced by person, product, and model - the kind of breakdown finance teams have been asking for as the bills climbed.

The controls go past reporting. An admin can set a spending ceiling for an entire workspace, carve out separate limits for individual teams, and grant exceptions to the people who genuinely need more. Workers can check their own usage against that budget and request more headroom with context attached, so an approval becomes a decision rather than a surprise. There are even APIs to pipe the numbers into a company’s own finance dashboards.

It’s telling that these controls now live inside the vendor’s own product. The era of handing teams an uncapped AI account is closing, and visibility-plus-limits is becoming table stakes. Uprovd’s view is that a spend cap is only half the job - the other half is tying each of those dollars to the outcome it bought, so finance can fund what works instead of merely throttling what doesn’t.