AI's Increasing Role in Fraud Prevention

In the few last years, artificial intelligence has become a hot topic due to its multiple business applications. In Air Europa, it is currently being used in several areas of business such as pricing, voice recognition, e-commerce consumer behaviour, product recommendation, and fraud prevention.

Specifically, in the area of fraud prevention, airlines have to deal with fraud attacks taking place almost continuously. Fraudsters have a 24-hour job to find out where a breach in security or gap in procedure is and take advantage of it. These attacks range from CEO fraud, through ransomware attacks to credit card fraud. Some of them can be prevented with a set of security protocols and tools. However, others—mainly due to the volume of transactions to be evaluated— exclusively need AI and more specifically, machine learning, which is a branch of artificial intelligence that basically provides algorithms to identify anomalies like fraud patterns.

There are several areas in an airline where machine learning algorithms can be applied to identify fraud patterns. In Air Europa, we consider two groups. The first group would be the most critical one, meaning that a relevant increase of fraud ratios can result in severe damage for the airline. Credit card fraud and identity theft would be the most relevant ones. 

"Since, Algorithms Need To Be Tuned Up, Their Results Go Through Human Manual Review and a Feedback Loop Is Created To Keep Improving the Algorithm"

Uncontrolled credit card fraud can result in critical cost issues for e-commerce and call centre channels. Moreover, it can result in penalties by the main card schemes and ultimately, in the vetoing of card acceptance, which would be effectively a point of no return for the airline. Identity theft can also result in penalties by the correspondent data protection authority. Moreover, it has a severe impact in terms of reputation.

Because of the criticality in Air Europa, we think that these two areas need to be protected by a fraud prevention specialist as the complexity of the algorithms and the resources required are high enough. Besides, these specialised companies are able to gather information from other airlines and companies and can foresee and stop credit card frauds based on all the information processed by them.

The second group would represent those areas where fraud attacks are rarer and thus, the impact of the attacks is lower—both in terms of costs and reputation. Some examples are baggage claim fraud and loyalty programs fraud.

These areas are so specific to airlines that there are no outsourcing options. So, Air Europa has adopted an in-house approach. A team of data scientists has been set up to confront these challenges by developing anomaly detection algorithms. Since, algorithms need to be tuned up, their results go through human manual review and a feedback loop is created to keep improving the algorithm.

In conclusion, Air Europa being an airline wants to be focused on its main business, so it cannot invest extensively as a high-tech company would, but it does not want either to miss the boat in this respect as we believe that artificial intelligence is going to be a top priority for successful business in the next decade. This strategy allows us to keep track of new developments in the area as well as maintain a reasonable level of in-house knowledge of the field.