How Virgin Trains is using ServiceNow to identify fraud during ticket refunds

Barj Duhra, ServiceNow Platform Owner for Virgin Trains

In 2018, travel and tourism were the third-most targeted industry by cybercriminals, after finance and retail, an IntSights study found. And travel scams account for nearly half, 46% of all fraudulent transactions, Experian reports. With 39 million passenger journeys covering 22.4 million miles a year, Virgin Trains is the biggest rail operator and files roughly 750,000 customer service cases annually. This includes complaints, praise and requests for refunds due to delayed service or cancellations. Between 10% and 30% of complaint compensation goes to fraudulent claims, analysts have found.

In February 2018, Virgin Trains implemented a new machine learning-powered fraud detection module, co-designed by ServiceNow and UK ServiceNow Partner Development Technology firm UP3, in its Customer Service Management platform. This replaced an industry-standard legacy system built on MS Dynamics.

The older technology did not have the ability to scan transactions for traditional scam indicators, such as unusually high numbers of refund requests from a single traveller, nor could it flag both good and bad customers for fraud. To determine which cases were real fraud and which were false positives, Virgin Trains carried out manual investigations to analyse it and create a report.

Meanwhile, honest customers waited for their refunds for weeks or months. By alienating customers, Virgin Trains lost even more revenue than the amount paid out for fraud itself. About 10% of cases were false positives—legitimate claims misidentified as fraud. False positives cost six times more than fraud. When you have a false positive, you lose the entire lifetime value of that client.

Dean Underwood, Head of IT Services, Virgin Trains.
Dean Underwood, Head of IT Services, Virgin Trains.

With its new fraud module, Virgin Trains can weigh claims by assessing and scoring hundreds of data points reflecting customer behaviour, such as payment method, IP and home address, email address, case history and more. Not only can the system detect impossible journeys, where a customer claims to be on two different trains at the same time, it can also uncover and visualise linkages between payment types, email accounts and more data to bust sophisticated fraud rings.

What Virgin Trains has now is a consumer dashboard that gives a visual relationship map. It shows not just one individual, but other people who may be linked. Virgin Trains calls it a data explosion—it brings it all together with just one click. The implementation of its new machine-learning fraud detection system has helped Virgin Trains reduce the complaint handling process with consistent, positive affect. All told, Virgin Trains has slashed the time it spends to reach complex case management decisions by 95%, down to 20 minutes on average.

The more data fed into the system, the smarter it becomes and the faster it can identify fraudsters. Individuals, organisations and domains tagged for fraud get added to the company’s red list and all future transactions associated with them are halted. This enables the company to move from detection to prevention.

Highlights 

  • Travel and tourism were the third-most targeted industry by cybercriminals.
  • When you have a false positive, you lose the entire lifetime value of that client.
  • The system can detect impossible journeys, where a customer claims to be on two different trains at the same time.
  • What Virgin Trains has now is a consumer dashboard that gives a visual relationship map. 
  • Virgin Trains calls it a data explosion—it brings it all together with just one click.
  • The more data fed into the system, the smarter it becomes and the faster it can identify fraudsters.
  • Virgin Trains has slashed the time it spends to reach complex case management decisions by 95%, down to 20 minutes on average.
Barj Duhra, ServiceNow Platform Owner for Virgin Trains
Barj Duhra, ServiceNow Platform Owner for Virgin Trains.