Take off with AI

For many travelers, flying can be a stressful and frustrating experience. Delays, cancellations, missed connections, lost luggage, long lines, and security checks are some of the common pain points that can ruin a trip. For airlines, a delayed flight can cost between 45 and 182 euros per minute due to the resulting expenses, such as missed connections,  provided hotel stays, passenger claims, and left-behind baggage.¹ 

Artificial intelligence (AI) has the potential to transform the aviation industry and make air travel more seamless, efficient, and thereby customer centric. Moreover, AI can help aviation executives achieve their operational goals: a faster, more stable and reliable turnaround, which, in turn, boosts customer satisfaction and loyalty. However, not all AI methods are the same, and different situations require different approaches. In this article, we will explore how AI can be used in three different ways – with hindsight, real time, and foresight analyses – in order to speed up air travel.

Increasing passenger satisfaction by improving punctuality

Many factors contribute to the overall passenger satisfaction. According to our long-standing experience with airline and airport customers, punctuality is the single biggest lever with an impact of over 20 percent in determining an airline’s passenger satisfaction.² Yet punctuality levels at major European airports are very low. For example, in the summer period of 2024, 40 percent of flights were delayed by more than 15 minutes.³ Therefore, operational stability and reliability are major areas of focus for aviation executives.

Achieving these goals is not easy. The aviation industry is complex and involves many interdependent factors and stakeholders. For a flight to take off on time, ground operations (e.g., fueling, loading, and catering), flight operations (e.g., crew rotation), terminal operations (e.g., bag drop, security checks, and gate operations) and maintenance (e.g., ramp checks) need to work seamlessly. This requires a lot of co-ordination between the airline, the airport and different service providers. In addition, the growing number of flights and theinflux of passengers, in combination with labor shortages of skilled personnel, means the complexity of airport operations has skyrocketed.

To ensure operational stability and reliability in such highly complex airport systems, data-driven and AI-based solutions are becoming essential. They can provide transparency for complicated operations, help in decision-making, automate repetitive processes, and aid in operational planning. 

Looking back: AI-based hindsight analysis

One of the ways that AI can improve the aviation industry is by providing hindsight analysis, which is the retrospective examination of past events and outcomes. This can help airports and airlines to identify the root causes of problems, such as delays, cancellations, or complaints, and to learn from their mistakes and successes.

For example, causal AI (or causal inference) is a method used to discover the causal relationships between many intricately related variables, such as the turnaround delay time in relation to weather conditions, air-traffic control, staff availability, equipment requirements, and individual process delays. By using causal AI, airports and airlines can understand how and, in particular, which of the hundreds of factors affect their total turnaround delay time. Such advanced AI methods make it possible to identify the impact of hidden drivers on turnaround delays. Using these insights and a deep knowledge of industry-specific operations, executives can take strategic decisions focusing on the processes with the greatest lever in avoiding future delays. For example, flights where a quick turnaround is crucial, executives may have to consider allowing dual boarding and re-prioritizing the cleaning process.

Natural language processing (NLP) is another valuable AI method for hindsight analysis. NLP is used to analyze and understand human language. This includes, for example, sentiment analysis and classification methods. By using NLP, airports and airlines can process and sort textual information, such as feedback and complaints from customers, and provide more proactive and personalized responses. This can help identify hidden pain points in operational processes, for which otherwise no data or only unstructured textual data exists. What’s more, analyzing safety reports using NLP methods enables the historic and real-time tracking of safety events at airports. This allows for quick intervention in case of common or crucial safety findings. In general, NLP can help make new data sources usable and provide a more holistic picture. 

Better decisions with AI-based real-time analysis

Another way that AI can improve the aviation industry is by providing real-time analysis, which is the simultaneous examination of current events and outcomes. This can help airports and airlines to monitor and manage their operations in real time, and to react quickly and effectively to changing situations and demands. AI-based real-time analysis and decision-support systems are useful for improving overall punctuality, especially with respect to standardized rerouting and replanning decisions. By using advanced optimization methods, airports and airlines can adjust their schedules and plans in real time, helping them to cope with unforeseen events, such as delays, cancellations, or disruptions. 

Unexpected inbound flight delays are a good example of where this would be needed. Firstly, passengers with connecting flights need to be rerouted. Planners needing to make complex decisions quickly should be supported by algorithms that take all relevant factors into account. There are advanced optimization methods that can handle multiple tasks – such as minimizing passenger delays while keeping the cost of rerouting low – all while respecting complex constraints, such as time frame and availability. Similarly, decisions such as choosing a new parking position for deboarding and reboarding passengers at the gate or apron, after an initial position is lost, can also be addressed with optimization methods. In critical situations, these AI-powered tools would require approval from the planner, or would need to produce a selection of possible solutions. In this regard, AI can give the person in charge transparency over the consequences of their actions, helping them to make more informed decisions even in complex situations. Being able to respond dynamically in real time to issues or delays via automated systems will give airlines and airports a cutting edge in improving stability, and therefore punctuality.

Telling the future with AI foresight analysis

A third way that AI can improve the aviation industry is by providing foresight analysis, which is the prospective examination or prediction of future events and outcomes. This can help airports and airlines to anticipate and prepare for the future, and to leverage the opportunities and challenges that may arise.

Foresight analysis is especially useful in the airline industry for aircraft maintenance activities. Every minute an aircraft is on the ground results in lost revenue, so it is crucial to minimize maintenance time. With many constraints, such as obligatory repetitive tasks, the available time between flights, man-hours and material availability, this is a challenging task even for routine maintenance. It becomes even harder when un-scheduled incidents such as defects in the cabin area or paint jobs on the airframe arise, which cannot be completed during the planned shift. This calls for planned buffer time for non-routine maintenance via advanced prediction models. Such models allow for dynamic buffer calculations for maintenance events, and they can be integrated into automated planning tools to ensure the continuous operability and punctuality of aircrafts.

In fact, AI-powered digital twins are at the forefront when it comes to making strategic use of foresight analytics as an AI method. Preparing for the future means developing contingency plans for potential future scenarios. This could include modelling scenarios for disruptive events, such as extreme weather, unexpected labor shortages, or resource blockages, and determining optimal countermeasures via an optimization- or simulation-backed digital twin. Similarly, strategic decisions or investments today should be based on simulation and sensitivity analyses of important questions such as: what happens if we increase or reduce the minimum connecting time? What is the impact on punctuality if we add staff to baggage handling? How must I adapt my maintenance plan to ensure airworthiness while dealing with a low availability of man-hours? Data-driven simulations can model these effects and help derive important strategic decisions and increase operational resilience.

Strategic use of AI for seamless air travel

AI is a major lever for airlines and airports to increasing punctuality and should be used to deliver insights that aid decision-making processes: from strategic decisions based on hindsight analysis (e.g., identifying and combating outbound delay causes) and real-time decisions (e.g., rerouting passengers) to foresight operational decisions (e.g., optimal maintenance planning).

Each of these areas of analysis meets urgent needs in the aviation industry, yet airlines and airports are still largely only focusing on understanding their problems with hindsight analytics, or they are firefighting problems based on real-time feedback. To remain competitive in the industry, players cannot afford to be reactive and solely learn from the past to understand the issues driving delays. Strategic countermeasures need to be identified to mitigate issues before they arise, and processes improved to prevent them altogether. AI-based methods allow for dynamic responses to real-time unexpected events or disruptions, and allow players to proactively engage before inevitable events occur. Aviation companies can therefore utilize AI to act more strategically, rather than solely reactively.

To this end, not only do airports and airlines need to make use of the varied AI methods and analyses available, but they also have to place more attention on proactively preventing issues before they arise. This can be achieved through continued strategic decision-making via AI-supported hindsight analysis, coupled with intelligent real-time and sophisticated foresight analysis. Ultimately, this will give executives the cutting edge in the industry and help them achieve their goal: making air travel as seamless as possible.