The Microsoft 365 outage: why one supplier shouldn't be able to stop your business

What happened, in Microsoft's own words

Problems began late on Thursday afternoon, UK time. Microsoft's first updates described "elevated load across multiple regions" and "a number of servers being unexpectedly unavailable", which put strain on the servers that were left. Engineers changed the way servers were health-checked, restarted the ones that had stopped responding and worked through a backlog of queued email. By Friday morning Microsoft said email queues were "approaching normal levels", and on Saturday it said it had "completed our efforts to address the underlying issue of the incident."

On Friday Microsoft gave its first explanation of the cause: "an issue during an earlier network capacity upgrade" that surfaced after a more recent network change. It apologised too: "We are deeply sorry for the disruption this incident has caused our customers and partners."

So was it an attack? On everything Microsoft has said so far, it may not have been malicious at all. Microsoft says there is "no indication" of malicious activity, and that the outage was not related to a cyberattack or to AI agents. That's Microsoft's early account, though, not the final word. Its full post-incident review hasn't been published yet, and we'll update this article when it is.

One more warning. Banks told customers to watch for scam emails and calls pretending to be about the outage. If a message asks you to "re-verify" your Microsoft account or pay to restore a service, don't click. Go to Microsoft through your usual route instead.

Who was affected

This wasn't just an inconvenience for office workers. In the UK alone, according to the Associated Press:

• NHS services were disrupted, and doctors had trouble emailing prescriptions to pharmacies. The National Pharmacy Association warned of medication delays.
• A trial in Manchester was halted and the jury sent home.
• Police couldn't take online crime reports and asked people to phone about urgent matters.
• The UK government said it was working with Microsoft and had contingency plans.

Around the world, the stock exchange operator Euronext told companies to phone in their trades, banks in Hong Kong reported disruption, Air India rerouted its email, and two listed companies delayed their results announcements. Even OpenAI and Anthropic reported problems connecting to Microsoft 365 tools.

Why so much more than email stopped

Most businesses don't think of Microsoft 365 as their booking system or their CRM. But over the years, a lot of everyday business software has been wired into it:

None of these tools are bad. The risk is that they all lean on the same platform, so when it falls over they fall over together, and there's nothing you can do but wait.

Could it happen again, now AI is moving into business systems?

Almost certainly, yes. This was the second major Microsoft 365 outage of 2026: in January, an outage dragged on for nearly ten hours, reported The Register. In July, an Azure outage in the US was traced to a bug in Microsoft's automated network maintenance system, which removed network routes from more devices than intended, according to Microsoft's preliminary review. And in July 2024, a faulty security update from CrowdStrike crashed what Microsoft estimated was 8.5 million Windows devices worldwide.

The big cloud platforms are extraordinarily reliable most of the time. But they're also enormous, and very concentrated. The UK competition regulator found that AWS and Microsoft each account for 30 to 40% of UK cloud spending. When one of them has a problem, a large share of the economy has it at the same time.

Now add AI. Microsoft says this outage had nothing to do with AI agents, and we've no reason to doubt that. The question is what comes next. Gartner predicted that 40% of enterprise applications would include task-specific AI agents by the end of 2026, up from less than 5% in 2025. Those agents are being given access to email, calendars, CRMs and finance systems, often all inside one supplier's cloud. Gartner also expects that by 2027, 40% of enterprises will scale back or switch off autonomous AI agents because of governance failures.

More automation means changes made faster than people can check them, more systems depending on each other, and more things that stop when one shared service does. The July Azure incident shows how an automated process can widen a small mistake. We're not saying an AI agent will cause the next big outage. We are saying that as AI is built into more of the same few platforms, the number of things that break at once when one of them fails is going up, not down.

Why a bespoke system keeps your business running

We build bespoke business software: booking systems, CRMs, job management, client portals and compliance systems, made around the way a business actually works. They don't depend on Microsoft to run:

• 0Microsoft accounts needed

  Your system has its own secure sign-in, so staff can log in even when Microsoft can't check their account.

• 1system, yours

  It isn't one tenant among hundreds of millions on a shared platform. It runs on our own infrastructure in UK data centres.

• £0per-seat fees

  You're not renting each user a licence, and your core operations don't hang on someone else's subscription.

In practice, that means that when Outlook goes down, your team can still sign in, look up a booking, update a job and see a customer's history. If you want, we'll send the system's own emails (confirmations, reminders, alerts) through a separate route from your office mailbox, so customers keep hearing from you too.

No honest developer will promise that any system never goes down. Ours run on servers too. The difference is that a problem affects one system rather than hundreds of millions of users at once, and fixing it is our team's whole job, not a ticket in a queue. You also own your system and your data, so you're never locked in.

You don't have to give up Microsoft 365 for email and documents, either. The aim is simply that the system your business runs on doesn't stop when your email does. If your core processes currently live in spreadsheets, Access databases or a tangle of connected apps, our legacy system conversion service moves them into one system you control. See what that looks like in our case studies, or read where your business data actually lives.

Six things to do before the next outage

Microsoft 365 outage: quick answers

Sources

• Associated Press: Microsoft says services fully restored after hourslong global outage (14 November 2026)
• CNBC: Microsoft apologizes for massive outage that took down Outlook and other 365 services (12 November 2026)
• The Register: Microsoft 365 outage drags on for nearly 10 hours (23 January 2026)
• BleepingComputer: Microsoft blames massive Microsoft 365 outage on maintenance bug (July 2026)
• Microsoft: Helping our customers through the CrowdStrike outage (20 July 2024)
• Competition and Markets Authority: Cloud services market investigation (final decision, July 2025)
• Gartner: 40% of enterprise apps will feature task-specific AI agents by 2026 (August 2025) and uniform governance across AI agents (May 2026)