An aircraft experiences a technical issue before its next flight. Maintenance checks it and determines it cannot fly as planned.
From there, the impact quickly spreads beyond maintenance. It affects aircraft assignment, crew scheduling, gate planning, connections, passenger communication, and the recovery process.
None of this changes the engineering decision. Safety always comes first, and it should.
But it raises a wider question: when something changes in one part of the operation, how fast can the rest of the airline understand and react?
For me, this is where resilience really shows. It’s not just about how well one part manages disruption, but how quickly the entire operation can coordinate its response.
I work on the other side of this system: passenger check-in, boarding, baggage handling, airport operations, and managing disruptions. Maintenance is not my area. And it’s often in passenger-facing disruptions that hidden operational issues become visible first. Every delay in information or decision eventually reaches passengers or frontline staff.
So the real challenge isn’t just within maintenance or operations. It is about how effectively information and decisions are shared between them.
Resilience Across System Boundaries
This is already shared information between maintenance and operations. An aircraft that becomes unavailable does not just sit in one system until someone notices.
Operations Control usually knows what to do next and decides whether to wait for the aircraft, assign another aircraft, delay the flight, cancel it, or change the overall schedule. However, one challenge that is not always easy to include in operational systems is uncertainty.
Operations naturally want an answer. When will the aircraft be available? Maintenance might have an estimate, but an early guess is not the same as a confirmed time for returning to service. In practice, however, operations often need a specific time even when maintenance can only provide an estimate. An estimate can quickly become a fixed time.
For example, if it is probably between ten and twelve, it might be considered eleven. From there, eleven may start to be seen as a commitment. This matters because decisions depend not only on the estimated time but also on how confident everyone is in that estimate. Waiting two hours when the estimate is stable can make sense. But waiting two hours when there is a chance it could be six hours late can lead to a very different situation. A more connected operation can track both the current expectation and how certain it is. That may sound like a small difference, but in a disrupted operation, it can have a significant impact.
Context Matters as Much as Connectivity
A technical situation needs to become an operational decision, and that decision must make sense across several systems, teams, and touchpoints. Not everyone needs the technical detail. A gate agent does not need the maintenance log. A passenger application does not need to understand the defect. But different parts of the operation do need the right context. An aircraft substitution, for example, is not just a registration change. It may mean a different seat map, different operational constraints, new crew considerations, and passenger-facing consequences. So the question is not simply whether systems are connected. In most airline environments, some level of connectivity and orchestration already exists.
The more important question is whether the right information reaches the next part of the operation with enough context to support a good decision. That is where resilience starts to become an ecosystem capability rather than a feature of one system.
Resilience and Modernisation Are Connected
Airlines operate technology environments that have evolved over many years. Maintenance systems, departure control systems (DCS), crew applications, airport platforms, and operational tools will coexist with newer Order-based standards and environments for quite some time.
Therefore, modernisation cannot simply mean replacing everything.
A more realistic approach is to keep what still works, separate capabilities where it makes sense, and introduce new ones gradually.
This also means building technology that is easier to connect and evolve, without creating new dependencies along the way. For example, if an airline introduces a new
disruption-management or passenger-communication capability, it should not need to rebuild every connection around DCS, crew, and airport systems to make it work.
That flexibility is part of resilience, too.
Airlines that can adapt one part of their technology landscape without disrupting the rest are not only easier to modernise but also better able to respond when operations do not go as planned.
AI Can Help Move From Prediction to Action
Much of the current discussion around AI in aviation focuses on prediction: failures, delays, missed connections, or capacity constraints. These capabilities can create real value, but prediction is only the beginning.
The more interesting question is what happens next. AI can help assess three things more quickly.
First, the operational impact: what happens to the next sectors, aircraft rotation, crew, or airport operations?
Second, the passenger impact: who is now at risk, what alternatives exist, and when does waiting create more disruption than recovering earlier?
Third, the response itself: which options are available, which decisions can be recommended or automated, and where human oversight remains essential?
The technical conditions needed for this type of orchestration are increasingly in place. Most modern systems offer APIs, and older systems are usually connected through integration layers.
So the question is no longer whether AI can access these systems, but rather what it can understand and do once it is there.
A plane’s status is just one piece of data. Making good operational decisions also depends on factors like downstream rotations, crew constraints, passenger connections, airport limitations, commercial priorities, and how confident we are in the maintenance estimate.
This is where AI becomes especially useful.
The goal is not simply to make individual systems smarter, but to help the entire operation understand trade-offs faster and respond more effectively.
One Operation, Not One System
The answer isn’t a single huge platform. Different systems will keep doing their own jobs well. Airlines will continue to work with several providers, and old and new technologies will coexist for years. The key point isn’t whether information moves through APIs, events, or other ways of connecting. It’s whether the right information can move quickly between systems and include enough context for the next step to act on it.
Maintenance focuses on maintenance, passenger systems on passengers, and operations control still owns the operational decisions. But when there’s a change, the airline should be able to respond as a single operation rather than as separate systems.
That is a more useful way to think about resilience.
Returning to the aircraft in our example, it’s still grounded, and the engineering decision remains the same.
What has changed is how the response is handled. Operations gets an earlier view of how long it might last and the level of uncertainty. A replacement aircraft can be considered alongside its broader impact. Seat changes are flagged before passengers find out on their own. At-risk connections become visible while alternatives are still possible. Airport teams receive an updated picture of the situation instead of trying to piece it together themselves.
None of this requires maintenance to become passenger operations, or passenger systems to understand engineering. It requires information, context, and decisions to work better across the operation.
As aviation invests more in predictive maintenance, AI, and increasingly intelligent systems, I think one question becomes more important:
When something changes in one part of the operation, how quickly can the rest of the airline understand it, decide on a course of action, and execute?
That may become one of the clearest measures of airline resilience.
