Microsoft 365 outage breaks ChatGPT, Claude connectors

ChatGPT and Claude stayed online on Thursday, Nov. 12, but users who had linked either AI assistant to Microsoft 365 found it could no longer read their email, calendars or files. A global Microsoft ↗ 0.36% outage took down Outlook, Teams and other workplace tools, and the connectors that OpenAI and Anthropic offer for those services failed with them.

The problem was in Microsoft's systems, not the AI platforms. Status updates from both companies showed that only their Microsoft integrations were affected.

What the AI companies reported

Anthropic first posted at 17:52 UTC, about 12:52 p.m. ET. It said requests to mail, SharePoint, Teams chat and calendar through Claude's Microsoft 365 connector might fail and that new connections might not complete, according to Claude's status page. Its later updates blamed "an upstream provider" and said Claude for Outlook was also affected, with some users unable to open the add-in. By 22:23 UTC, Anthropic named the cause directly as "a Microsoft 365 service incident." It said mail, calendar and search tools were "still failing for most requests," and that the incident affected claude.ai and Claude Cowork.status.claude.com

At 1:29 p.m. ET, OpenAI said it was investigating "an issue with Microsoft connectors where new connections may be broken and tools are not working as expected," according to its status page. Reuters Thomson Reuters Corporation ↗ 0.39% reported that the outage was "also hitting connectors used by OpenAI's and Anthropic's AI services." Chosun Biz noted that neither chatbot went down entirely, but features such as searching email or pulling up schedules and documents stopped working.fidelity.com+2

A long outage

Microsoft services were disrupted for about 10 hours on Thursday, and the outage hit businesses and government agencies around the world, Reuters reported. Outlook problem reports on Downdetector peaked above 130,000 in the U.S. at around 1 p.m. ET. Microsoft said features that depended on Exchange Online and email were affected, while services that did not rely on Exchange Online kept working.fidelity.com

Microsoft at first blamed heavy load on its infrastructure. It later said "a number of servers being unexpectedly unavailable" were "straining the remaining infrastructure," ABC13 reported. The company paused all deployments and redistributed traffic, according to BleepingComputer. Microsoft executive Charles Lamanna said there was no sign of a cyberattack, Malwarebytes reported.abc13.com+2

Ripple effects

The disruption spread into Asia-Pacific early on Friday, local time. Australia's spy agency and federal police were among the organizations affected, and the U.S. House of Representatives said its users were hit, Reuters reported. Workhorse Group moved its third-quarter earnings release to Nov. 16 because of the outage.fidelity.com

Australian Cyber Security Minister Tony Burke said Microsoft believed the situation was in some ways "similar to what we saw with respect to CrowdStrike" ↗ 0.35%, "where there is an update that has gone awry."