OpenAI's ChatGPT Work and Microsoft Outlook both went down Monday
Two of the busiest tools in the enterprise stack — Microsoft's email backbone and OpenAI's enterprise agent — spent Monday afternoon degraded at the same time. There's no reported connection between the two failures, and that coincidence is arguably the more interesting part.
OpenAI's ChatGPT Work stopped running tasks for roughly three hours on Monday, and Microsoft's Exchange Online degraded inside the same window — two outages at two companies that are simultaneously partners and rivals, with no indication either caused the other.
OpenAI said on its status page that users across multiple subscription plans "may be unable to start or continue tasks in ChatGPT Work," and that it was "continuing work on implementing a mitigation." The company's ChatGPT account put it more plainly on X: "ChatGPT Work isn't working right now. Sorry about that, we're working on a fix!" Microsoft confirmed through its 365 status account that "users may experience degraded functionality with various Exchange Online functionalities," adding that it was reviewing service telemetry and diagnostic data to isolate the source.
User reports tracked by Downdetector had already peaked earlier in the day — close to 4,600 for Outlook and more than 1,100 for OpenAI products as of mid-afternoon ET. Neither company has said what caused its incident.
The timing is a coincidence, but the shape of the failure isn't. ChatGPT Work, introduced in July, is OpenAI's answer to Anthropic's Claude Cowork: an agent built to take on multi-step projects — polished slides, web apps, sheets, documents — rather than answer questions. When a chatbot goes down you lose a conversation. When a task agent goes down mid-run you lose work that was in flight, and the failure mode is an interrupted process rather than a slow answer. That is a materially higher reliability bar than a chat window has ever had to clear, and it's the one every enterprise agent is now being measured against.
For Microsoft the incident is smaller but pointed. The company that put the first $1 billion into OpenAI in 2019 — a stake that grew to $13 billion — now runs a productivity suite that competes with its old partner on several fronts, and on Monday both halves of that relationship were broken at once.
What to watch: OpenAI's post-incident writeup. If the root cause turns out to be capacity rather than a bad deploy, the economics of the enterprise agent story get more expensive before they get better.
If your company moved real workflows onto an AI agent, how much downtime would you tolerate before you pulled them back? Tell us in the comments.
