Ing. Calogero Bono
On November 13 Claude, the assistant from Anthropic, went through a period of high error rates. The problems affected claude.ai, Claude Code and the developer console, with trouble logging in and using the service. According to the official dashboard cited by Punto Informatico, Anthropic reported a problem with the Microsoft 365 integration, while the status page mentioned an investigation into authentication issues on Claude.ai and Claude Code, at a time when the Microsoft services blackout was still ongoing. It matters because more and more businesses are building their daily work on AI tools hosted by third parties.
A few terms before we get into the facts. Down or outage means a service that does not respond or responds poorly. API is the interface through which software and third-party apps talk to Claude. Connector is the link that lets Claude read data from other services, such as Microsoft 365. Status page is the official page where the company reports incidents and the progress of the fix. Downdetector is a portal that collects user reports, useful but not official.
What happened to Claude on November 13?
Anthropic confirmed a high number of errors. This is reported by Punto Informatico, which cites the official dashboard, and confirmed by Android Authority, according to which the outage affected claude.ai, the Claude API, Claude Code and the connectors.
Users had difficulty using the chatbot and accessing the service. Again according to Punto Informatico, Downdetector showed a spike with thousands of reports, which then dropped in the late morning.
How did the status page describe the incident?
The Claude status page shows the classic stages. First an investigation into an authentication problem on Claude.ai and Claude Code. Then the rollout of a fix, with monitoring of recovery. Finally the resolution of the problem affecting claude.ai and Claude Cowork.
The incident involved claude.ai, the Claude Console (platform.claude.com), Claude Code and Claude Cowork. Android Authority adds that Anthropic saw recovery on claude.ai and the API, but warned that the errors could return while it worked on the solution.
What does the Microsoft blackout have to do with it?
According to Android Authority, the first problems appeared after the Microsoft failure, which caused the Microsoft 365 connector to fail for most requests. Punto Informatico notes that Microsoft 365 and Outlook had been down since a long blackout that began the day before.
The sources also say that a separate problem emerged during the day. So there is no single confirmed explanation, and the status page refers to authentication, not Microsoft.
Which Claude tools were involved?
The scope is broad. It ranges from the web app to developer tools like Claude Code, and to the console for those who integrate Claude into their own products. Over recent months Anthropic has extended the reach of the service, as we covered with the office AI suite and the link to bank accounts in Personal Finance. The more functions run through Claude, the more an outage is felt.
Limits to know
We do not know the precise technical cause. The sources do not give the total duration, the number of users affected, or whether the errors came back after the fix. Downdetector data are user reports, not official measurements. In the sources provided, Anthropic has not published a post-incident analysis.
What changes for those who work digitally
If your work runs through Claude, an hour of errors is an hour of slowed activity. Anyone using the API inside a management system, a website or an app should plan clear messages for users and automatic retries. Anyone using connectors to Microsoft 365 should remember that the chain has several links, and one broken link is enough. It is also worth knowing where to look, namely the status page, before wasting time hunting for a fault in your own system.
Our take
A few hours of downtime says nothing about the quality of Claude, and the same goes for Microsoft. It does say something about how we build our processes. Too many companies entrust critical tasks to a single AI provider with no alternatives. The comparison between providers, like the one in our piece on Claude and OpenAI, should also cover service continuity, not just price and response quality.
Frequently asked questions
Is Claude working again?
Yes, according to the status page the problem on claude.ai and Claude Cowork has been resolved. Android Authority however reported that the errors could return while the fix was being rolled out.
Which services were hit by the Claude outage?
claude.ai, the Claude Console, Claude Code and Claude Cowork were affected. Android Authority also mentions the API and the connectors.
Was the Claude outage caused by Microsoft?
This is not fully confirmed. Anthropic pointed to a problem with the Microsoft 365 integration, but the status page also mentions authentication.
Where can I check Claude's status?
On the official page status.claude.com, which lists incidents and resolution stages. Downdetector instead collects user reports and is not an official source.
How long did the outage last?
The sources do not say precisely. Punto Informatico reports the situation improving at 11:45 on November 13, after a peak in reports.
What to watch after the Claude outage
The open questions now are whether Anthropic will explain the cause and whether the errors will return. For you, what matters most is how exposed you are to a single provider.
- Check whether Anthropic publishes an incident report with the technical cause.
- Map the workflows that depend on Claude and Microsoft 365 together.
- Subscribe to status page alerts so you know before your clients do.
