Microsoft 365 Outage Hits Copilot and AI Connectors Worldwide

A wide outage of Microsoft 365 services began on the evening of Thursday 12 November 2026. It hit the Outlook email service, Teams and the Copilot artificial intelligence assistant, according to the BBC. According to the same outlet, user reports of Outlook problems peaked at nearly 140,000 worldwide around 18:00 GMT, as measured by the outage-tracking site Downdetector.

Quick definitions: Microsoft 365 is an online suite of workplace tools used by companies, including Outlook, Teams and OneDrive, as the BBC describes it. Copilot is Microsoft's AI platform. Downdetector collects user reports, so its figures count reports, not the real number of people affected, as the site AITnews explains.

What has Microsoft said about the cause?

According to CNN Arabic, Microsoft announced on X that it was investigating reports of problems, that part of its infrastructure was under increased pressure across multiple regions, and that it had redistributed traffic across its global infrastructure to balance that pressure.

The company then explained, per CNN Arabic, that the high pressure did not come from an increase in traffic but from a number of servers going out of service unexpectedly, which put a burden on the remaining infrastructure. The BBC reports almost the same thing: servers that were unexpectedly unavailable strained the other systems.

Microsoft later said, according to CNN Arabic, that part of the affected servers had come back and that it was rolling out the change more widely, while preparing automatic recovery systems to continuously detect unresponsive servers and restart them. According to the BBC, it also warned that some restored services can become unresponsive again and that customers in any region could be affected.

How did the outage spread, and which services did it hit?

• Geographic spread: users in the US, Japan and Singapore reported disruptions, according to the BBC.
• Start: the German site Golem says the outage reached users in Germany from Thursday evening and that Microsoft spoke of users in all regions possibly being affected. According to Golem, the investigation first covered Exchange Online mailboxes that were unreachable in North America, then it was said that connectivity problems could affect other regions, and the problems spread to the Asia-Pacific region on Friday.
• Services: Golem, citing Dr. Windows, names Teams, OneDrive, Copilot, Microsoft 365 and the Microsoft Store.
• Functions: according to the BBC, Microsoft said users might be unable to view some files or details in their calendar, while messaging and online meetings were largely functional.

Was it a cyberattack? What is established and what is not

The BBC says it understands there is no indication that the outage was caused by a cyber-attack, and AITnews writes that the available evidence does not show the incident was caused by an attack. That does not mean the root cause is known.

According to AITnews, updates on the two incidents logged under the identifiers EX1550543 and MO1550821 show that Microsoft examined a temporal correlation with a network configuration change, as part of its diagnosis and rollback of possible changes. But, as the same site points out, a temporal correlation is not enough to declare a root cause. So the only explanation the company has published, as far as we read, is servers going out of service; why they went out of service is not established.

Figures differ between sources. The BBC speaks of about 140,000 reports worldwide at the peak, Golem says Downdetector at times logged more than 130,000 reports in the United States, and AITnews, citing Reuters, mentions about 138,000 reports. These are reports, not counts of affected users.

Why did AI tools stop working?

The BBC wrote that Copilot might not work properly and that OpenAI and Anthropic were hit because they work with Microsoft's systems. According to the BBC, OpenAI said it was investigating an issue affecting its connections to Microsoft's services, which caused some tools not to work as expected.

AITnews explains that the status updates at OpenAI and Anthropic pointed to problems in the Microsoft 365 connectors, which let some users access their mail and calendar data or search company information. A connector is a component that links an AI tool to another service so it can read that service's data. The site insists on the distinction between a disrupted external connector and a stoppage of the core ChatGPT and Claude services: those alerts do not allow anyone to say the two platforms went down worldwide.

Also according to AITnews, Zoom reported degradation of some of its integrations with the Microsoft ecosystem, covering calendar, OneDrive and Zoom Scheduler; Notion reported effects on some Outlook integrations; and Veeam reported backup and restore problems for Microsoft 365. As for delays in email, notifications and verification codes at GitHub and Okta, their coinciding with the outage is not enough to claim a single cause unless both companies confirm it. No reliable evidence allows describing what happened as a general collapse of the internet either.

What does the outage reveal about cloud dependence?

The BBC notes that Microsoft reported an outage of its Azure cloud platform last month, which affected its productivity tools and companies worldwide. It also notes that the electric truck maker Workhorse Group, which was due to report third-quarter results on 12 November, postponed them to 16 November, citing business disruptions caused by the outage.

For AITnews, the incident reveals the interdependence of digital services: enterprise applications rely on email, calendar, authentication and programming interfaces at other providers. The effect can show up as undelivered meeting invitations, delayed messages, failed scheduled backups or an AI agent unable to reach a user's mail and calendar. The site specifies that these examples describe the nature of the functions exposed, not a list of confirmed cases in every organisation.

What this changes for a company in the Gulf and the Arab world

What follows is the newsroom's reading, not facts attributed to the sources, apart from the first two points, which AITnews recommends. Any Arab company that relies on Microsoft 365 can:

• check the Microsoft service status page and the service health dashboard in the Microsoft 365 admin centre before concluding that service is back to normal, because remediation updates were still coming in on the morning of Friday 13 November, UAE time;
• check the status page of each external service to find out whether the outage lies in the integration with Microsoft or in the core service;
• list the AI agents that read mail and calendar through connectors, the first exposed in a comparable incident, and plan a manual fallback for meetings and important messages.

To go further on that last point, see AI agent governance in the enterprise.

What we still do not know

According to AITnews, the overall root cause of the incident is not definitively established, nor is the total number of users and organisations affected or its worldwide financial cost. It is also not confirmed that all the failures logged at other providers were part of the same incident. Since remediation updates were still arriving on the morning of 13 November, it would be wrong to conclude that all services are back to normal: the experience can vary by region, subscription type, service and integration used.