How artificial intelligence is transforming revenue management for airlines

At Amadeus, we’re heavily invested in AI and machine learning(ML) in many of our solutions, including revenue management (RM). We’re also driving innovation in this area with participation at key industry events like AGIFORSand our partnerships with universities to foster PhD researchers. These elements are enabling our airline partners to embrace the amount of data they have and use it to make informed business decisions. As airlines service more passengers with different profiles, while offering more travel options and more types of services and ways to combine them, AI and ML backed solutions will be key in helping them make better decisions.

AI already powers revenue management for airlines

Forecasting future passenger demand has always been one of the most challenging tasks for revenue managementsystems. This is not only because of the technical complexity but also because of the industry’s fast-changing requirements. 

Where once upon a time, airlines only needed to predict how many people might be traveling from point A to point B in one single cabin, now there are so many more variables to consider. What was once just an economy and a first-class cabin has transformed into economy, premium economy, business, first and even more in-between. Then there are the extras - suitcases, priority seating, extra legroom - each airline can only provide so many of these services, and if they want to maximize their revenues, they need to sell each extra kilo of baggage and each seat with extra legroom at the best possible price for any given flight. 

Indeed, as airlines develop bigger networks, Amadeus' AI powered revenue management system has evolved from focusing on capacity utilization - by forecasting each flight’s number of passengers, no-shows and cancellations - to more advanced forecasting of travelers’ price sensitivity in different markets, network-wide passenger flows as well as what drives the way travelers choose between different options (customer choice models). 

Towards the holy grail of revenue management

At Amadeus, we’re pushing the frontiers of revenue management with the development of a Competitor-Aware Revenue Management System (CARMS). Earlier this year at AGIFORS,our Chief Scientist, Thomas Fiig, presented on the concepts and feasibility of CARMS.His team demonstrated that this innovative vision is practically attainable given advancements in AI/ML technology and available data. They showed that it is possible to predict future competitor price changes with a Mean Absolute Percentage Error (MAPE) of about 10% - which is similar to the accuracy that airlines can predict their own prices. This is a remarkable achievement and paves the way for major breakthroughs in revenue management for airlines.

The proof is in the pudding for airlines

So, does AI/ML technology make a difference to bottom lines? Just ask Singapore Airlines.With Amadeus systems underpinning many of Singapore Airlines customer touchpoints, the carrier was able to use Amadeus´ AI powered systems to better understand its customers and to use Amadeus “Willingness To Pay estimation” algorithms to better adjust availability, and thus pricing, according to customer demand. This has, in part, helped Singapore Airlines grow in passenger flown revenues in 2018 – a significant improvement which exceeded initial estimates. While Singapore Airlines’ flown revenue improved, so did staff productivity, thanks to the automation of many operational activities that previously required manual intervention. 

What´s next for AI/ML in revenue management for airlines?

Amadeus continues to be at the forefront of the AI revolution: AI was a key component of the Amadeus Altéa Revenue Management solutions from the start and continues to be incorporated into the Amadeus Airline Platform. Today, Amadeus´ Offer Management, Dynamic Pricing, Advanced Merchandising solutions, and Customer Experience Management all use AI modeling to help airlines adopt more open and agile ways of working so that passengers can have a more personalized and rich travel experience. Tomorrow, thanks to Reinforcement Learning techniques, our solutions could even be able to learn by themselves how to adapt to passengers and market evolution or to detect exceptional events and adjust accordingly. 

We’re also contributing to the greater field of AI research through a collaboration with the University of Nice Sophia-Antipolis in the form of research studies conducted by a PhD student under the supervision of Professor Jean-Charles Regin.

Finally, to support and accelerate our progress, Amadeus R&D teams can now also count on the Machine Learning Servicesteam. Their mission is to support, educate and provide technical tools, and to foster the usage of artificial intelligence and machine learning techniques across all our solutions.

And based on our accolades at AGIFORS, our strategy seems to be working. Stay tuned for more about how we’re using AI and ML technology to help airlines achieve better business results.

Editor’s note: This post is part of a series on artificial intelligence and machine learning technology for airlines. Have a look at the first post here.