OpenAI investigates Microsoft 365 connector errors as wider outage continues
Update, November 13 at 3:08pm Sydney: separate Sites publishing issue
A fresh check of OpenAI Status also found a separate incident labelled "Sites: Publishing failing". OpenAI listed that incident as identified, with elevated errors for the impacted services and work underway on a mitigation. The Microsoft 365 connector incident remained under investigation. The status notice did not establish that the two incidents shared a cause, or that all ChatGPT conversations were unavailable. No completed mitigation or recovery time was confirmed in the notice checked.
OpenAI is investigating errors affecting the Microsoft 365 connector in ChatGPT. Its official status page said new Microsoft connections may be broken and connector tools may not work as expected when checked on Friday, November 13, Sydney time.
An integration problem, not a confirmed platform-wide shutdown
The incident was listed as investigating and affecting ChatGPT. That scope matters: a problem accessing Microsoft information through a connector is not evidence that every ChatGPT conversation or every OpenAI API is unavailable. The status notice did not provide a confirmed repair time or a completed root-cause finding.
Separately, UptimeRobot reported that its external check could reach chatgpt.com without unusual response times or error codes. The displayed check was at 03:00 UTC on November 13, equivalent to 2pm in Sydney. This supports a narrower conclusion that the website was reachable from that monitoring location. It does not test every signed-in conversation, subscription tier, connector or region, and is not an all-clear for affected Microsoft integrations.
The Microsoft outage context
The Published reporting indicated that the wider Microsoft 365 disruption affected Outlook, Teams and Copilot, and also had effects on AI companies working with Microsoft's systems. It specifically cited OpenAI's investigation into connections to Microsoft services. Microsoft attributed its broader connectivity problems to servers becoming unexpectedly unavailable, increasing pressure on remaining systems, according to the the published report. This is the reported context, not an independently established technical explanation for every OpenAI connector error.
Microsoft's public status page still displayed service degradation for Microsoft 365 when checked. It separately said no Azure issues were detected at that time. The current incident should therefore not be described as proof that all Azure services or all AI platforms were down. Our earlier Microsoft 365 report covers the wider disruption and its reported Australian impact.
Service conditions can change after these checks. The photograph is a general network-cabling illustration, not an image of OpenAI or Microsoft infrastructure, a diagnostic screenshot or evidence of the fault.
The published report