June 2018, London
The bot that helps you find a spot on the train
Commuters using Busybot, a feature embedded in the Trainline app, may be more likely to get a seat on the journey home this Christmas. Senay Boztas reports
There are seven core problems everyone faces when travelling by rail in Europe, according to train and coach booking and information site Trainline. One of them is finding a carriage with a seat. While some train companies invest in apps to weigh the physical carriages and find which is lightest, Trainline had a simpler solution: ask the travellers.
And travellers, it turns out, were obliging. Jonathan Moore, Trainline's chief product officer, told the EyeforTravel Amsterdam 2017 conference: “BusyBot was to answer a simple question: are there any free seats on your carriage? A train operating company was talking about spending hundreds of thousands on pressure gauges – and we had to tell them we had solved it free with software.”
BusyBot, which is found with the journey tracker part of Trainline’s app, asks customers to report how busy trains are and uses the hundreds of thousands of reports to make predictions for any service on any route.
What Trainline discovered, was that customers were happy to do the work for the firm in building a dataset if they saw clear benefit. “We see 150,000 unique inputs every week into this feature,” he said. “There’s a real selflessness in this number – there’s no reason to interact with this feature unless you make someone else’s life a bit better on another day.”
53% of customers are willing to share personal data for tools that help them make decisions
Source: Columbia Business School
The circumstances in which users are willing to hand over share data is one that also came up in a presentation from Oleksandr Medovoi, President & CEO, AltexSoft, an EyeforTravel sponsor. He shared research from Columbia Business School, which found that 53% of customers are willing to share personal data for tools that help them make decisions.
For Medovoi then, companies would be wise to focus closely on what users really want, with dedicated user behaviour analysis (UBA). Noting that companies could “no longer survive using standard practices,” Medovoi claimed firms that UBA techniques could see engagement rise by 2.5 times and conversions by 80%.
A world first
Luckily for Trainline, it operates in an environment where online distribution is still emerging, so users are fairly easy to please if you get it right. BusyBot is clearly a winner with customers, and Moore says the company is committed to building AI driven customer solutions that people actually want.
Another recent example is that it applied artificial intelligence to crunch a huge data set and build a model that not only shows how much ticket prices increase nearer the day of travel, but to estimate precisely when they will rise and by how much.
Without humans there wouldn’t be machines
“This is a real world first,” Moore said. “We focused on the dirty little secret of rail – the price only goes in one direction 99% of the time and that’s up. That Chester to London ticket, which would have cost £23 pounds, on the day of departure will cost £126.
“Our iOS app now has price prediction, after a data science team’s months of work. This was released two months ago – purely data driven – and had never been done before.”
Answering the question of whether travel will ever pit machines versus humanity, Moore’s answer was a no.
“Without humans there wouldn’t be machines,” he said. “At Trainline, we innovate with people at the heart.”