Who’s in Charge When the Systems Go Down? What Microsoft’s and the FAA’s Outages Teach Every Team

By Priya, Research, HUMAN HIVE

Who is in charge when the systems your team depends on go down? Whoever you named before it happened. Microsoft 365 went down on Thursday, November 12, 2026, and the FAA stopped every departure in the country on Friday. Neither one is blamed on AI, and both show that the plan for a bad day is a people plan: who decides, who talks, and how the work keeps moving without the tool.

The 60-second version

  • Two big system failures in two days: Microsoft 365 on November 12, and a four-hour nationwide FAA ground stop on November 13.

  • Uptime Institute says failures to follow established procedures are still the leading cause of outages tied to human error.

  • This week, name who makes the call when a key tool goes dark, and run a 15-minute "tool is down" drill with your team.

What happened

Thursday, November 12: Microsoft. Starting around 18:00 UTC, The Register reported, Exchange Online, Teams, Copilot 365 and Defender for Office 365 became unavailable for many customers. On X, Microsoft blamed "a problem that's causing servers to fail automated health checks and become unavailable." One detail matters for every leader: the admin tools Intune and the Microsoft 365 Apps Admin Center went down too, which made it harder to fix.

SecurityWeek reported that, by Microsoft's account, a process that sets up secure connections became overwhelmed and servers stopped handling requests. A recent network configuration change happened when the trouble started, and Microsoft began rolling it back. Microsoft said, "At this time, we have no evidence of malicious activity associated with this incident," and promised a Post-Incident Report with the root cause (incident EX1550543).

Friday, November 13: the FAA. The FAA ordered a nationwide ground stop at 3:54 PM Eastern and lifted it at 8:15 PM, Al Jazeera and Reuters reported. It was the first nationwide ground stop since January 2023. FlightAware data reported by CBS News counted more than 10,000 flights delayed and at least 4,000 cancelled. CNBC reported that the FAA cited a "temporary telecommunications outage" affecting multiple Air Route Traffic Control Centers and halted flights out of an "abundance of caution." Transportation Secretary Sean Duffy said, "Our switches, our routers actually got overloaded with data." Homeland Security Secretary Markwayne Mullin said it "appears this was mechanically driven." According to Cirium data cited by CNBC, about 11,000 flights were scheduled to leave U.S. airports after 4 PM Eastern that day, carrying 1.5 million passengers.

Why it matters (the human side)

Neither failure is an AI story. Both are dependency stories, and AI is about to make every team more dependent on systems than ever.

Look at what each organization did in the first hour. The FAA had a rule ready: when communication is in doubt, stop the planes. Somebody owned that call, made it, and explained it. Travelers had a miserable night, and that's the trade a clear rule buys: an inconvenience instead of a risk. Microsoft had the harder problem, because the tools engineers use to diagnose and manage the system were caught in the same failure.

Most teams are somewhere in between. They know the tool is essential. Almost none of them have written down who decides, what work continues by hand, and how customers hear from a human while the screens are dark.

The data backs that up. Uptime Institute's Annual Outage Analysis 2026 (May 13, 2026) found that "failures to follow established procedures remain the leading driver of human error-related outages," and that "inconsistent or unclear processes are also common." In Uptime's 2025 survey, 57% of respondents said their most recent major outage cost more than $100,000, and for the second year running 1 in 5 reported costs over $1 million. Andy Lawrence of Uptime Intelligence put it plainly: digital infrastructure is "remarkably resilient," but "further resiliency gains are becoming harder to achieve." The gains still on the table sit with people.

What Mitch has taught on this

Mitch has been saying for a while that AI is turning into infrastructure. In his talk on AI for destination teams, he said, "AI now is really needing to be compared to more electricity or the smartphone or computers. It's no longer an ROI experiment." And to every team whose early AI projects broke: "You're just learning how to integrate a new type of electricity into your system."

When something becomes electricity, you need a plan for the blackout. Mitch's line from the same talk is the right frame for it: "AI can clear the friction, but the human gets to land the moment." The system carries the volume. People carry the judgment, the calm and the customer when the volume stops.

Best practices: build your "systems down" plan

  1. Name the call owner for each critical tool. For email, CRM, your AI assistant and your phone system, write one name and one backup who can declare "we're in outage mode." The FAA's stop worked because the decision had an owner.

  2. Keep a second channel that doesn't share the first one's failure. If your team chat and your email run on the same platform, they go down together. Keep a phone tree or text list that works without both.

  3. Write the manual version of your top three workflows. One page each: how we take an order, answer a customer and hand off a shift with no software. If nobody can write it, nobody can do it.

  4. Keep admin and recovery access outside the blast radius. Microsoft's admin tools went down with its services. Ask IT: if our main platform fails, can we still get into the tools that fix it?

  5. Script the first customer message now. Two sentences, approved in advance, sent by a named person within 30 minutes. People forgive outages. They don't forgive silence.

What's driving people crazy this week

Copilot 365 went down with Exchange and Teams on November 12 (The Register). For teams that now draft, summarize and search with an AI assistant built into the same platform, one outage took the work and the helper at the same time. The human fix: know which of your daily tasks depend on the assistant, and keep a plain way to do the two that matter most.

What this means for the next 1 to 3 years

As AI agents take on more steps in a workflow, an outage won't only mean "I can't read my email." It will mean work that was moving on its own stops in the middle. We expect three shifts:

  • "Manual mode" becomes a skill you train, the way pilots practice flying without the autopilot. Teams will rehearse a day without the tools the same way they rehearse a fire drill.

  • Every agent gets a human owner and a stop button. The question in every rollout review shifts from "what can it do?" to "who stops it, and what do we do then?"

  • Calm becomes a competitive advantage. When everyone's systems wobble at once, customers remember the company whose people stayed clear, kind and in charge.

How we evolve

The human skill to build here is deciding under uncertainty. As systems get smarter, the moments they fail get stranger and harder to read, and someone still has to make a good call with half the facts. That's the heart of Mitch's ADAPT thinking and of Amplify, Not Just Automate: let the machine carry the load, and make sure the humans can still steer.

One small daily practice: at the end of each day, pick one tool you used and ask, "If this disappeared tomorrow morning, what would I do first?" Say the answer out loud in one sentence. A month of that and your team has a recovery plan nobody had to schedule a meeting for.

Try this today

  • Name the owners (10 minutes): list your five most critical tools and write one call owner and one backup next to each.

  • Ask at your next team meeting (5 minutes): "If email and chat went down at 9 AM tomorrow, how would we reach each other and our customers by 9:30?" Write down every answer.

  • Draft the outage message (10 minutes): two sentences you could send customers in an outage. Get it approved now, not during.

Free tool: the "systems down" planning prompt

"Here are the five tools my team uses every day and what we use each one for: [list]. For each tool, help me write: who should decide we are in outage mode, the simplest manual way to keep the most important work going, how we contact each other if this tool and our chat are both down, and a two-sentence message to customers. Ask me questions wherever you are guessing about how my team works."

FAQ

What caused the Microsoft 365 outage on November 12, 2026?
Microsoft said servers were failing automated health checks and that a process that sets up secure connections became overwhelmed. A recent network configuration change happened around the time the trouble started, and Microsoft began rolling it back. Microsoft said it had no evidence of malicious activity and promised a Post-Incident Report with the root cause.

Why did the FAA stop all flights on November 13, 2026?
The FAA cited a temporary telecommunications outage affecting multiple Air Route Traffic Control Centers and halted departures out of an abundance of caution. The ground stop ran from 3:54 PM to 8:15 PM Eastern, the first nationwide stop since January 2023.

Were either of these outages caused by AI?
No. Neither was blamed on AI. The lesson is about dependence: as AI becomes part of every system, teams need a human plan for when systems fail, with named decision owners, a backup channel and manual versions of key work.

Where HIVE can help

Leaders decide with half the facts every day. Practice it together, live, before it counts with a Decision Heist for your team. Rolling AI into more of your work? Our AI enablement work builds the human side of the plan with it.

Get The HIVE Daily Buzz in your inbox.

Previous
Previous

The AI Capability Paradox: Why Senior Leaders Win With AI and Junior Staff Fall Behind

Next
Next

The Bench Test: What Nvidia’s Overnight CEO Change Teaches Every Team About Succession