Published on 16 November 2026

The global disruption of Microsoft services has highlighted a reality that many organisations have yet to fully acknowledge: having advanced technology does not guarantee the ability to continue operating when that technology fails. The rapid expansion of artificial intelligence adds a new dimension to this challenge.

During 12 and 13 November, millions of people around the world experienced difficulties using tools that form part of their everyday activities. Outlook, Teams, SharePoint and Copilot, among other Microsoft services, suffered disruptions that affected businesses, public administrations, educational institutions and healthcare organisations.

What initially appeared to be just another technology incident ultimately became an issue of international significance. Companies were unable to communicate normally, professionals lost access to their emails, and organisations were forced to rely on alternative procedures in order to continue operating.

The scale of the incident cannot be explained solely by the number of users affected. Its true significance lies in the extent to which we have turned certain technology platforms into essential components of modern society.

Microsoft has explained that the root cause was a technical issue associated with a previous network capacity expansion, which became apparent following a recent change to its infrastructure. The situation resulted in an overload of certain systems and complicated the recovery of services. The company has also ruled out the possibility that the incident was caused by either a cyberattack or artificial intelligence systems.

This clarification is important because it allows fact to be separated from the speculation that has circulated in recent hours. From a cybersecurity perspective, however, the conclusion is far broader: organisations do not need to suffer a cyberattack to experience consequences similar to those one could cause.

 

When Technology Is No Longer Available

For years, cybersecurity has primarily been associated with preventing unauthorised access to systems, protecting information from theft and safeguarding data integrity. While this remains necessary, it is no longer sufficient. Security is also about ensuring that services remain available when they are needed.

A company may have robust protection against intrusions and still find itself paralysed because a technology provider ceases to function. It may maintain comprehensive backups only to discover that it has no viable alternative means of communicating with employees. It may even possess sophisticated protection tools yet temporarily lose the ability to use them because they depend on the same affected infrastructure.

The Microsoft outage reminds us that availability is an integral part of security and that both must be addressed together. This highlights a concept that deserves a much more prominent place in business decision-making: cyber resilience.

Being resilient does not mean operating systems that never fail. That is impossible. It means having the ability to anticipate problems, withstand their effects, maintain essential activities and restore normal operations within an acceptable timeframe. The distinction may appear subtle, but it fundamentally changes how technology risks are managed.

 

The Problem of Over-Reliance on a Small Number of Providers

One of the debates reignited by the incident concerns the concentration of digital services among a relatively small number of large technology companies.

The adoption of cloud platforms has enabled organisations of all sizes to access capabilities that, until recently, were available only to large enterprises. We have gained efficiency, scalability and enhanced opportunities for collaboration. The challenge arises when multiple critical processes depend, directly or indirectly, on the same infrastructure.

An organisation may use different applications, services and providers, only to discover that they all share a common technological dependency. Under such circumstances, what appears to be a localised failure can trigger consequences on a much broader scale.

This does not mean abandoning cloud services or questioning the benefits offered by major technology providers. It means understanding our dependencies and assessing what would happen if one of those services became unavailable for several hours or even days.

Business leaders should ask themselves a simple question: could my organisation continue carrying out its essential activities if its primary technology platform ceased functioning tomorrow? Many organisations would probably struggle to answer because they have never tested this scenario.

 

Artificial Intelligence Introduces a New Dependency

The incident also coincides with a particularly significant moment in technological transformation. Artificial intelligence is no longer merely a tool for retrieving information, drafting documents or generating content. Increasingly, AI solutions are being integrated into business applications, accessing corporate information, executing processes and carrying out actions with varying degrees of autonomy.

So-called AI agents represent an important step in this evolution. Consider an organisation that uses agents to manage customer requests, access documentation, prepare commercial operations or coordinate internal processes.

What happens if those agents depend on services that suddenly become unavailable? What becomes of the tasks they were performing? How can we determine which activities were completed, which remain pending and which might be repeated once connectivity is restored?

These questions have practical implications. An agent that temporarily loses access to an application may leave an operation unfinished. If the system resumes operation and automatically retries the task, it could duplicate an action if suitable controls are not in place. In administrative, commercial or financial processes, such situations can create errors that are not always easy to identify.

It is worth stressing one point: there is no evidence that artificial intelligence caused the Microsoft outage. Microsoft itself has explicitly ruled this out. The relevance of AI lies in the lessons we can learn from the incident and in how we design the new systems we are introducing into our organisations.

The more autonomy we grant to technology, the more important it becomes to have mechanisms in place to supervise its operation, halt processes when necessary and recover operations without compromising data integrity.

Artificial intelligence can help improve resilience, detect anomalies and accelerate incident response. However, it can also introduce new dependencies and amplify the impact of failures if implemented without an appropriate strategy.

 

Automation Does Not Mean Surrendering Control

There is a growing tendency to measure the success of digital transformation by the number of automated processes or the volume of tasks that can be performed without human intervention. This is an incomplete perspective.

A truly advanced organisation is not necessarily the one that automates the most, but the one that knows how to retain control over its processes, even when automated systems fail to behave as expected.

This requires careful attention to matters that do not always receive the prominence they deserve. Artificial intelligence systems must incorporate oversight mechanisms, records that provide visibility of actions taken and procedures for managing errors or interruptions.

It is also necessary to define which decisions can be executed automatically, which require human validation and which activities must be halted when sufficient guarantees are not available.

Restoring a service should not simply involve confirming that it is operational again. Organisations must verify that interrupted processes have been completed correctly and that information remains consistent and reliable.

In certain environments, particularly those involving essential services, industrial infrastructures, healthcare or financial systems, these precautions become even more important. We cannot allow the adoption of artificial intelligence to create processes that nobody knows how to supervise or recover when problems arise.

 

Cybersecurity Must Also Protect Us from Failures Nobody Caused

Another important aspect of this incident is the reaction that often occurs when a major technology platform stops working. The first question is usually whether a cyberattack is to blame.

The concern is understandable. Digital threats are part of our reality, and their consequences can be severe. However, focusing all our attention on attacks can cause us to overlook other risks capable of producing disruptions of comparable impact.

A configuration error, infrastructure failure, defective update or poorly managed technological dependency can all compromise organisational continuity. From a business perspective, what matters is not only what caused the disruption, but how it affects our ability to continue delivering services.

For this reason, cybersecurity strategies must evolve towards models that integrate prevention, detection, response, business continuity and recovery.

It is equally important to review how third-party risks are managed. Procuring a technology service does not mean fully transferring responsibility for the consequences of its unavailability. Providers have obligations and responsibilities, but every organisation must understand which processes depend on them and implement measures proportionate to their importance.

 

What Should We Learn from What Happened?

The Microsoft outage should prompt us to revisit a number of assumptions that are often taken for granted.

First, we need to identify our genuine technological dependencies. It is not enough to know which applications we use. We must understand which services they rely upon and what consequences would arise if those services were interrupted.

Second, continuity plans should be reviewed. Having a document that describes how to respond to an incident does not guarantee that the organisation is actually capable of doing so. Procedures must be tested, particularly in scenarios where email, collaboration tools and certain security services fail simultaneously.

Third, organisations should maintain independent operational alternatives. If the emergency communication channel depends on the same platform that has become unavailable, it can hardly be considered a genuine continuity solution.

Fourth, resilience should be embedded into the design of artificial intelligence projects. Before automating a critical process, organisations should understand how to stop it, recover its execution, verify completed actions and determine who will assume control if the technology fails.

Finally, organisational leadership must be actively involved. Technology continuity cannot be regarded solely as the responsibility of IT or cybersecurity departments. Its consequences affect customers, employees, operations, revenue and reputation.

 

A Reflection That Goes Beyond Microsoft

Digital transformation has enabled us to achieve things that would have seemed unimaginable only a decade ago. We can work from anywhere, share information in real time and increasingly deploy intelligent systems capable of helping us make decisions and carry out complex tasks. This represents remarkable progress.

At the same time, however, we are building organisations whose ability to function depends on an increasingly interconnected network of technology services.

The Microsoft outage should not be interpreted as a reason to distrust technology. That would be the wrong conclusion. Instead, it should remind us that no infrastructure, however advanced, is entirely free from failure.

Artificial intelligence will make this reflection even more important. We are moving from using digital tools to entrusting them with part of the execution of our processes. This evolution requires a greater capacity for oversight, control and recovery.

For years, one of cybersecurity’s central questions has been how to prevent someone from bringing our systems to a halt. Perhaps it is time to give equal importance to another question: what will we do when our systems stop, even if nobody has attacked them?

The answer will largely determine which organisations are genuinely prepared for the years ahead. The true measure of digital maturity is not how dependent we are on technology when everything works, but how effectively we can continue operating when it does not.