When digital collaboration faltered, an old standby helped keep business moving: the telephone. But for CIOs, the bigger lesson wasn't about old-school technology — it was about the dependencies their continuity plans can miss.
The Microsoft 365 outage that began Nov. 12 lasted more than a day and affected users around the world. Some organizations whose own services were working still struggled to reach customers and partners who relied on Microsoft.
Beginning Nov. 12, Microsoft's service health updates reported that users were unable to connect to Exchange Online mailboxes. Related problems affected some features in Teams, SharePoint Online, Copilot and Defender XDR. Microsoft's preliminary findings traced the failure to a latent networking-gateway defect exposed by traffic shifts and reduced capacity.
Requests queued until affected servers ran out of processing capacity. Microsoft said the incident was a technical issue, not a cyberattack or the result of AI agents.
The usual response to an outage this size is to call it a wake-up call. But after repeated incidents, that truism can fall on deaf ears.
"Every big outage gets called a wake-up call, and then everyone goes back to sleep, because the takeaway becomes review vendor risk, which you could write after any outage," said Nora Jones, vice president of product development at PagerDuty and founder of Jeli. "And nobody was asleep at the wheel."
When Microsoft went down, phones were picked up
Because Microsoft services are so widespread, the outage affected more than its direct customers. Companies that depended on people using Microsoft platforms found those partners and customers harder to reach.
"Fortunately, we’re not a Microsoft shop so we did not lose email nor group collaboration capability," said Brian Greenberg, CIO and CTO, senior partner and strategic adviser at Fortium Partners. "However, many clients and partners were affected."
Greenberg said his organization used phone calls to reconnect with affected contacts and make other arrangements. Some people also used LinkedIn, text messages, Zoom or Slack when available. For Greenberg, the simplest approach was picking up the phone.
"The interesting thing I noticed is the general reluctance to pick up the phone for many," Greenberg said. "I don’t mean texting, which is also another method that works, I mean picking up the phone and calling and speaking with the person to discuss in real time, live, synchronously."
Greenberg wasn’t the only one picking up the phone. At DigiCert, Eric Porter, senior vice president of finance and IT, did the same. He said his organization uses Slack extensively, so employees could stay connected internally during the outage. The bigger challenge was communicating with external partners who relied on Microsoft 365.
"We shifted to phone calls and text messages to keep those conversations moving until service was restored," Porter said.
“You can be fully up and still be stuck waiting on someone who’s down.”

Sharad Kumar
CEO and co-founder, FluidCloud
Sharad Kumar, CEO and co-founder of FluidCloud, also turned to the phone, even though his organization did not rely on Microsoft for all communications.
"A lot of our partner and customer work runs through Teams channels and shared calendars, and those conversations went quiet," Kumar said. "We moved anything time-sensitive to phone and other channels from our side."
The lesson for Kumar was that continuity is only as strong as the least resilient tool in your counterpart’s stack. "You can be fully up and still be stuck waiting on someone who’s down," he said.
The surprise wasn't the outage
For Dion Hinchcliffe, chief research officer at Frontier Research, the Microsoft outage itself was not a surprise.
"The surprise wasn’t that Microsoft went down," Hinchcliffe said. "It was how difficult recovery became as a shared infrastructure failure cascaded across services."
Beyond email, some features in Teams, SharePoint Online, Copilot, Defender XDR, Intune, Microsoft Graph and Windows 365 were affected. The scope differed across products and features; that does not mean every service was entirely unavailable. Jones noted that the affected list included tools many organizations use to manage incidents.
"If you coordinate a response in Teams, keep your runbooks in SharePoint and get your security alerts from Defender, all three were on Microsoft's affected list," Jones said.
Copilot and AI-enabled workflows that depend on Microsoft 365 services were another potential point of disruption. Jones noted that AI agents that triage tickets or draft replies through those services can be interrupted when the underlying systems fail.
"A lot of people were working with AI that day and lost it at the same moment they lost their email," she said.
The cascading failure caught some organizations off guard. Kumar said many companies have an inventory of their applications but lack a map of what those applications depend on.
That points to a resilience risk: Continuity planning often stops at the edge of a company's own systems.
"The dependency most companies overlook is the vendor layer they don't control," said Joe Locandro, CIO at Rimini Street.
“The goal isn’t zero downtime. It’s zero catastrophic downtime.”

Dion Hinchcliffe
Chief research officer, Frontier Research
Where to draw the line on downtime
Maintaining a second platform might sound like a straightforward answer to the question of resilience. In reality, it is neither simple nor cheap.
"Duplicating Microsoft 365 wholesale would be enormously expensive, operationally complex, and probably counterproductive for most enterprises," Hinchcliffe said.
Hinchcliffe said he once priced a full backup of essential systems for a financial services firm and called the result "ruinously expensive." That points toward a narrower goal.
"The goal isn’t zero downtime," Hinchcliffe said. "It's zero catastrophic downtime."
Greenberg said a business impact analysis weighs the risk and cost of losing each function, helping determine how much to invest in recovery. High costs or risks from even a short outage can justify a redundant system; lower-stakes functions might be able to wait for recovery. The analysis should involve the rest of the leadership team.
Greenberg added that faster recovery costs more and that chasing sub-second or minute-level redundancy can be a poor investment when a brief outage does little business harm.
That said, riding out an outage is easier to defend when everyone is hit at once.
"For better or worse, most of the world seems to be running some — if not all — services on SaaS platforms and there aren’t that many of them, so when there’s a major outage like we just had with Microsoft, everyone knows about it and we’re all in the same boat, more-or-less," Greenberg said. "So there’s greater understanding to the situation."
What CIOs can do now
The Microsoft outage was not the first to disrupt enterprise IT, and it will not be the last. CIOs can take several steps now to reduce the impact of the next failure.
Review your own day. Jones said teams should start with what people actually did during the outage. “People were wide awake and improvising, and what they improvised is the part worth studying," she said. "So review your own day, even though the outage was Microsoft's."
Set up channels outside the platform. Microsoft’s business continuity guidance recommends establishing and testing alternative communications channels. Locandro said paging tools can create text and email groups that require acknowledgment on personal phones and outside email addresses, with escalation to a second-in-command if no one responds. The channel also has to be independent. "The key question is whether your backup options depend on the same services as your primary tools," Porter said.
Identity is a common blind spot. Kumar said a problem in the identity layer reaches well beyond email. “When the identity layer has problems, it's not just email," Kumar noted. "It's file access, single sign-on into other apps, and now AI assistants that sit on the same data.”
That affects backup planning too. Kumar said a backup channel that signs in through the same identity provider “isn’t really a backup. It’s a second door into the same locked room.”
Keep files and contacts reachable. Some organizations continuously download email and files for offline access or copy them to another repository, Locandro said, although such approaches can be costly and difficult to manage. Porter said essential contacts should be accessible outside the affected platform.
Run the drill. A realistic test starts with losing the collaboration suite in the middle of an urgent decision. "Can people reach one another, access the information they need and communicate with customers?" Porter asked.
Sean Michael Kerner is an IT consultant, technology enthusiast and tinkerer. He has pulled Token Ring, configured NetWare and been known to compile his own Linux kernel. He consults with industry and media organizations on technology issues.