Artificial intelligence is rapidly changing the way the rail industry forecasts passenger demand and manages fares. But while the technology can process vast amounts of data and identify patterns at a speed no human could match, one revenue management expert believes there is a limit to how far the industry should automate its decision-making.
Eric Nordling, General Manager of Expretio, has spent more than 40 years working in revenue management across aviation and passenger rail. His view is that AI has an important role to play in the future of rail pricing – but it should be used to augment human expertise rather than replace it.
Eric Nordling, General Manager at Expretio, said:AI and advanced algorithms are incredibly useful for finding patterns in huge volumes of data. But they don't replace the experience of someone who has spent years working with demand, pricing and passenger behaviour. You still need people who can look at the results and understand whether they make commercial sense.
The distinction is increasingly important as rail operators look to make greater use of sophisticated forecasting technology.
Revenue management has always been based on predicting future demand. Operators need to understand how many people are likely to travel, when they will book, where they will want to go and what they are prepared to pay. The better those predictions, the better an operator can allocate its finite supply of seats.
The technology available to do this has changed dramatically.
When Nordling began working in the industry, forecasting systems might have relied on fewer than ten algorithms. Modern systems can use hundreds of models, drawing on years of historical data and continuously adapting their predictions as new information becomes available.
Eric Nordling, General Manager at Expretio, said:The computing power available today is on another level. We can look at far more information and test many more possible outcomes. That gives revenue managers a much stronger basis for making decisions than they had in the past.
But more data and greater computing power do not automatically mean better decisions. Nordling is particularly cautious about the idea that AI should be allowed to operate without human oversight.
Eric Nordling, General Manager at Expretio, said:The dangerous assumption is that you can simply tell an AI system to make the decisions and leave it to get on with it. There still needs to be an experienced person in the loop, checking what the system is telling you and making sure nothing has gone wrong.
That is particularly relevant in rail, where unusual events can have a significant impact on passenger demand. Sporting events, concerts, engineering works, disruption or changes to connecting services can all alter travel patterns in ways that historical data alone may not fully capture.
Human experience can provide the context that an algorithm lacks.
A revenue manager who understands the network may recognise that an apparent shift in demand is temporary, identify an anomaly in the data or understand why a particular recommendation does not fit the wider commercial picture.
Eric Nordling, General Manager at Expretio, said:The best results come from combining the strengths of both. Let the technology do what it does best – process information, identify patterns and forecast demand – but have experienced people supervising it and applying their judgement.
There is another misconception that Nordling believes the industry needs to address: that the most valuable technology is necessarily the technology that reacts fastest.
Rail operators are increasingly interested in real-time data, but Nordling argues that revenue management is fundamentally about looking ahead.
Eric Nordling, General Manager at Expretio, said:Revenue management isn't really about reacting to what's happening right now. It's about predicting what is going to happen next and making the right decision before demand arrives. If you wait until the train is filling up, you've already lost some of your opportunity to optimise it.
That distinction could become increasingly important as the UK’s rail industry considers how technology can support commercial reform.
Real-time information will undoubtedly have a role, but the real value comes from using data to anticipate passenger behaviour rather than simply responding to it. A system that can forecast demand across hundreds of possible itineraries gives operators the opportunity to influence travel patterns before passengers make their decisions.
The result could be a more sophisticated approach to fares, with discounts used to stimulate demand where trains have spare capacity and pricing adjusted to reflect genuinely high demand.
For Nordling, therefore, the future isn’t about choosing between AI and human expertise. It is about getting the balance right.
Eric Nordling, General Manager at Expretio, said:I don't see AI replacing the revenue management professional. I see it making that person more productive. The technology can deal with an extraordinary amount of information, but the human being is still there to challenge it, interpret it and make sure the outcome is right.
As rail operators increasingly turn to AI and advanced analytics, that may be the most important lesson of all. The technology can make the railway smarter, but it is experienced people who will ultimately determine how intelligently it is used.
This article was originally published by Expretio.
