Short on time? Read the key takeaways:
- The use of AI and machine learning in financial services has become increasingly important due to the shift towards online transactions and banking during the "new normal."
- According to a survey, bank card fraud is the second highest concern among consumers, and 49% of respondents do not trust their financial institution to alert them of suspicious activity on their accounts.
- The use of machine learning to combat fraud is not uniform across financial institutions, with smaller institutions having a lower adoption rate.
- A proactive approach to fraud detection using AI and machine learning can increase employee productivity and customer trust, leading to better customer experiences.
2021 marked the first year of “the new normal” and all the many significant life changes it introduced. Those include the digital banking transformation as more people needed to work and bank from home. Cloud computing helped function as a catalyst.
At the same time, as more of our lives, transactions and ways of connecting moved online, the security of all these interactions has become increasingly important. It’s safe to say the financial services industry has entered the era of artificial intelligence (AI).
The Unisys Security Index™ surveyed 11,000 people across 11 countries on a wide range of issues, including personal, financial, national and internet security concerns. Of the eight areas of concern measured, the 2021 results show that bankcard fraud ranks number two, up four points when compared to 2020 results. More concerning is that 49% of respondents do not trust their financial institution to alert them of suspicious activity on their accounts.
AI in financial services
What is AI? Artificial intelligence (AI) is the science of using computers or machines to solve problems that humans usually do with our natural intelligence. Machine learning is a subfield of artificial intelligence. At the most basic level, machine learning recognizes patterns in data and uses it to predict future behavior. Machine learning uses algorithms that iteratively learn from data, so computers find insights without being programmed to do so.
The financial services industry has long been technology-dependent and data-intensive, and data-enabled AI technology can drive innovation further and faster than we ever imagined. AI helps improve efficiency, enable growth, manage risk, and, most importantly, make a world of difference in customer experiences.
Once expensive, AI systems are becoming the norm, with costs and barriers to adoption falling. We see financial institutions making investments in cloud, big data and data applications such as microservices, eliminating the capital investment needed to develop, deploy and scale AI solutions. One way financial institutions have embraced AI in their customer service operations is via chatbots and conversational banking.
Machine learning is a critical fraud prevention tool
Surprisingly, machine learning adoption to combat fraud is not uniform across financial institutions. As consumers, we tend to assume our financial institution has the latest fraud prevention technology, no matter its size. Among major financial institutions in the U.S. with more than $100 billion in assets, more than 72% use machine learning systems to combat bank card and credit card fraud.
But what about smaller institutions with $10 million to $100 million in assets? It turns out that only 6% of those institutions in the U.S. have adopted machine learning systems to combat fraud, according to PYMTS. com research. Now, we see why nearly half of the survey respondents do not trust their financial institution to alert them to suspicious activity.
Many of these smaller financial institutions need a trusted partner to help them adopt machine learning technology and enable real-time monitoring. This will also decrease costs, as it’s much cheaper to prevent a fraudulent transaction rather than investigate a potentially fraudulent transaction after the fact.
Credit card fraud tops the list when we look at the most prevalent types of fraud. This type of fraud is also one of the most preventable. For example, financial institutions train their machine learning systems to detect credit card fraud by looking for patterns of typical suspicious behavior based on vast amounts of data.
The key word here is data. Financial institutions require vast amounts of fraud data to train machine learning systems. This data includes meaningful features of the credit card user’s transactions, including date, amount, product category, provider and user behavioral patterns. This data then goes through a trained machine-learning model that finds patterns and rules to classify whether a transaction is legitimate or fraudulent.
However, this is based on known types of fraud. What about new types of fraud that bad actors haven’t designed yet? To combat this, some financial institutions are creating synthetic data on unique fraud patterns to train their machine-learning systems. This is essential to moving from a reactive to a proactive approach to fraud detection. Synthetic data can augment a financial institution’s existing data set of fraud behaviors to improve its overall fraud detection machine learning models.
Why do people not trust their financial institutions?
According to PYMNTS.com, “Human analyst teams not only have a hard time detecting fraud as it happens – 45% of financial institutions say investigations take too long to complete.” These investigations, unfortunately, have high false-positive rates of up to 90%, where legitimate transactions are identified as being fraudulent. The result is frustrated customers, and worst-case scenario, customer abandonment.
Be proactive to combat fraud
Financial institutions that move from a reactive to a proactive approach to fraud detection will be better positioned to build trust with their customers. They’ll also stay a step ahead of rapidly evolving and increasingly complex cyber threats. To do that, financial institutions must embrace technologies like AI and machine learning to drive better customer experiences and combat fraud systemically. Doing so will increase employee productivity and strengthen consumer trust, which leads to greater loyalty and happier customers.