Dear Microsoft, Copilot has some notes

Microsoft had an outage. So I asked Copilot what it thought of the response

There is a special place in hell for communications professionals who publicly pick apart another company’s crisis management response without knowing the internal battles fought over every step of it. But I’d argue it’s fair game to ask Microsoft’s own AI service to critique the company’s response to the global outage that began on 12 November. So, I did.

(For the avoidance of doubt, Copilot’s responses are italicised while I opine in a lovely shade of orange.)

Dear Copilot,
Imagine you’re the corporate affairs director of a global technology company that suffers a major outage, disrupting email services for millions of customers across 70 countries for the best part of two days. What would your first statement say?
Kind regards,
Helen

The first release needs to do very little, very well: say what has happened, make clear you know how serious it is, establish that fixing it is the priority, and promise the next update. At this stage, I would resist explaining the cause unless I was absolutely certain.

What I would actually issue:

Service disruption affecting email access

We are currently experiencing a significant service disruption that is preventing customers in multiple markets from accessing email.

Our engineering teams are working urgently to restore service. We recognise how important email is to our customers, businesses and organisations around the world, and we apologise for the disruption this is causing.

Our immediate priority is to restore access safely and as quickly as possible. We are investigating the cause of the outage and will provide further information once that investigation has established the facts.

We will provide our next update by [time], or sooner if there is a significant development.

ENDS

In a crisis like this, the most important sentence may be the last one. Giving a specific time for the next update creates an information rhythm and stops the organisation disappearing into the technical response.

That’s interesting, Copilot. Would it surprise you to learn that Microsoft did not issue any press release? Beyond a handful of apologetic posts on X, its detailed communications took the form of technical updates on its Service Health Dashboard, available to Microsoft 365 administrators.

There is nothing unusual about Microsoft using its Service Health Dashboard to give administrators detailed operational information. What is unusual is appearing to treat that as sufficient communication for a global, prolonged outage affecting a service on which businesses depend.

Microsoft seems to have communicated primarily for administrators, despite having a problem of interest to users, customers, the media, investors, regulators and other stakeholders.

Okay, Copilot, here is Microsoft’s Service Health report on the outage. What do you make of it?

User impact:
Users may have been unable to connect to their mailboxes in Exchange Online through multiple connection methods.

More info:
Users may have been unable to connect to their mailboxes in Exchange Online through multiple connection methods.

Potentially impacted connection methods included:
- Outlook on the web
- Messaging API (MAPI)
- Exchange ActiveSync (EAS)
- Representational State Transfer (REST) API

Additionally, on-prem mail flow may have been affected if it flowed to or from Exchange Online.

Impact to additional Microsoft 365 services is described under MO1550821 in the Service Health Dashboard.

Scope of impact:
This issue may have affected users attempting to access the affected Microsoft 365 services.

Preliminary root cause:
While our analysis of the underlying root cause is ongoing, we’ve determined that this incident was caused by a failure in the systems responsible for establishing connections to the Exchange Online service. Specifically, traffic shifts during standard network maintenance and upgrades exposed a latent bug causing a rapidly increasing number of requests to queue for processing. As this accumulated, the affected servers exhausted their available processing capacity and could no longer accept new connections, resulting in widespread service access failures.

 

1. It reads like an incident report, not a public response

It has clearly been designed for administrators. ‘User impact’, ‘scope of impact’ and ‘preliminary root cause’ make perfect sense within a service health dashboard. But for a major global outage, it doesn't provide the broader public communication you would expect alongside that technical information.

2. It describes the technology better than the human consequences

We learn about MAPI, EAS and REST APIs, but there is no acknowledgement of what being unable to access email means to customers and businesses.

That doesn't require emotional breast-beating. Something as simple as:

We recognise the significant disruption this has caused customers and organisations that rely on Exchange Online.

would change the tone substantially.

3. But there's no obvious ownership or apology

That's what strikes me most from a corporate affairs perspective.

Microsoft tells us what the systems did, but not really what Microsoft says about what happened. There's no apology, acknowledgement of the impact, or much sense of the company speaking directly to its customers.

It is an explanation of an incident rather than a response to one.

4. Most importantly, this may be perfectly adequate for its intended purpose

And I think this is potentially the key to your article.

If Microsoft intended this only as a technical report for Microsoft 365 administrators, I'd be much less critical of it. Administrators probably do want references to MAPI, EAS, REST APIs and mail flow.

The corporate affairs question is therefore not:

Why was Microsoft's technical bulletin so technical?

It's:

Why, during a global outage, was this technical bulletin apparently expected to do the job of public communication as well?

That, to me, is the much stronger criticism.

And it makes your opening conceit with Copilot more interesting too. You aren't really asking AI to make fun of Microsoft's jargon. You're asking Microsoft's own AI whether, if it were the corporate affairs director, it would regard an administrator notice as an adequate corporate response to a major reputational event.

My answer is no.

As the outage continued, Microsoft’s administrator updates became considerably less penetrable, referring to changes that had not worked as expected and lapsing into corporate gobbledegook, such as ‘contact-processing component’. So, I had one final question for Copilot.

Without being too modest Copilot, do you find it strange that a company whose AI service is used by communicators around the world to help draft press releases and statements should communicate an incident of this scale in such a way?

Yes. And in this particular case, I find it quite striking.

Not because AI should be running crisis communications. It absolutely shouldn't. But this is almost the ideal low-risk use of it: take an accurate but highly technical operational update and ask, ‘Can an ordinary customer understand this?’

The underlying facts still have to come from the engineers. Humans still have to approve the message.

The failure isn't that Microsoft needed AI to write the statement. The failure is that nobody appears to have translated the technical response into a human one.

That's precisely where corporate affairs should add value. AI simply makes the absence of that translation harder to excuse, because producing a first plain-English draft now takes seconds.

AI can't solve that organisational problem. But it can remove ‘we didn't have time to rewrite it’ as much of an excuse.

Thanks Copilot. My work here is done!