Microsoft’s outage exposed the fragility of business workflows built on Exchange and Outlook, while OpenAI faced a separate enterprise-AI disruption. Microsoft focused on isolating and repairing an authentication failure; OpenAI said it was working to blunt errors and latency in ChatGPT Work.

Reports of Outlook trouble began arriving around 11:30 a.m. ET Monday, with users encountering email delays and failures, mailbox-search problems, authentication errors and other Exchange Online disruptions. By early afternoon, thousands of users had flagged problems on Downdetector, underscoring the breadth of an outage that hit the cloud service businesses use for email, calendars and contacts.¹

Microsoft initially said it had confirmed “degraded functionality with various Exchange Online functionalities” and was reviewing telemetry and diagnostic data to find the source.² The company later identified an issue in an authentication component, tested a remediation on part of its infrastructure and said the same component was affecting services beyond Exchange Online.²

By 5 p.m. ET, the fix was still not fully deployed. Microsoft said a misconfiguration might be stopping authentication components from deploying as expected, prompting a review of recent service changes.²

In parallel, OpenAI said ChatGPT Work had logged elevated errors and latency for roughly three hours. “Users across multiple subscription plans may be unable to start or continue tasks in ChatGPT Work,” the company said, adding that it was implementing mitigation.¹ OpenAI’s ChatGPT account put it more plainly: “ChatGPT Work isn't working right now.”¹

The simultaneous failures involved longtime partners whose relationship has also become increasingly competitive. But there were no reports linking the two outages: one centered on Microsoft’s authentication infrastructure, the other on OpenAI’s workplace AI agent.