- Trust & Safety
How is AI transforming fraud detection in banks?
This online shift in the industry has meant banks and other financial institutions are facing new challenges when it comes to creating safe and secure experiences for their customers. A report from TransUnion found that globally, online fraud attempt rates for financial services rose 149% between Q4 of 2020 and Q1 of 2021 alone.
Reducing fraud is top-of-mind for banks not only because it directly affects their bottom line — according to Javelin Research, total fraud losses climbed to $56 billion in 2020, a whopping 83% gain from a year earlier — but because it can damage their reputation, cause bad user experiences and harm customer retention.
Common types of banking fraud
The types of fraud identified in the financial services industry are varied. Below are a few of the most common types of banking fraud and their impact:
- Unauthorized transactions: Banking or credit card transactions that an account holder didn’t make or approve continue to be a nuisance to both banks and consumers alike. A Forbes article reported that roughly eight in 10 mobile banking users are concerned about credit card fraud. Further, the value of fraudulent transactions made with payment cards worldwide in 2021 was projected by Statistica to amount to more than $32 billion, a figure that could increase to $38.5 billion by 2027.
- Phishing scams: In its 2020 Internet Crime Report, the FBI reported that Americans lost more than $54 million in phishing scams that year. Both consumers and corporate employees can fall victim to phishing scams that can lead to unauthorized transactions, account takeovers (ATO), data breaches or identify theft.
- Identity theft: Reported as the most common type of complaint lodged by consumers by the FTC, identity theft has a major impact on both consumers and financial institutions. In 2020 alone, total financial losses from identity fraud were around $13 billion, according to results from Javelin’s 2021 Identity Fraud Survey.
But let’s not start waving the white flag to surrender to the online scammers just yet. Banks have a secret weapon at their disposal that can be used to help identify and prevent fraud: artificial intelligence (AI).
How are banks using AI for fraud detection?
Banks are finding that AI for fraud detection is fast, effective and efficient. In 2021, Fintech News reported that financial institutions are deploying AI-based systems in record numbers, with more than $217 billion spent on AI applications to help prevent fraud and assess risk. Even more promising is that 64% of financial institutions believe AI can get ahead of fraud before it happens.
There are a number of different applications of AI for fraud detection in the financial services industry. Analyzing transactions is one of the fundamental functions. From risk scoring to grouping consumers into identifiable clusters or “profiles,” each application is essential to building a robust fraud detection strategy. Here are some of the leading ways banks are using AI for fraud detection:
- Building purchase profiles: In order to accurately detect fraud, financial institutions must first understand what typical customer behavior looks like. Using machine learning to sort through vast amounts of data from past financial and non-financial transactions, banks are able to build and slot customers into a number of different profiles. Profiles are useful in that they provide an up-to-date picture of activity on an account and can help make predictions on future behavior. For example, an account could be profiled as “eats at restaurants on weekends,” “makes regular quarterly trips to Paris” or “fills their car up with gas after work.” A single account could be placed in hundreds of different profiles based on their activity, with the profile being updated in real time after each transaction. As transactions are made, AI determines whether or not it fits a pattern or departs enough from the norm to warrant being flagged.
- Developing fraud scores: All transactions can be assigned a fraud score by using data from past legitimate transactions, incidences of fraud and risk parameters set by the financial institution. The score, which takes into account variables such as transaction amount, time, card use frequency, IP address of a purchase, and much more, is used to assesses the fraud risk involved with that particular transaction. Fraud scores are used to either automatically approve a transaction, flag it for review or reject it altogether. Using machine learning, the accuracy of fraud scores improves over time.
- Fraud investigation: Machine learning algorithms can analyze hundreds of thousands of transactions per second. Neural networks take this capability a step further by making decisions in real time. These technologies are successful in culling the unmanageable number of flagged transactions that occur, and providing a concise list of those that require further investigation by a human counterpart. Investigating and prosecuting fraud claims can be incredibly time-consuming, so ensuring agents are armed with the proper tools to increase efficiency is essential. This application of augmented intelligence can help teams prioritize and streamline investigations.
- __Know Your Customer (KYC):__ AI-backed KYC measures can verify ID and documentation, match fingerprints and even perform facial recognition almost instantaneously. This powerful tool strikes the right balance between customer security and convenience.
Consumers continue to look to financial institutions to provide the ability to bank on the go and access their information online. At the same time, they expect their bank to create an environment where these transactions can be completed in a safe and secure manner. Working with an experienced Trust, Safety & Security partner like TELUS International can help financial services brands develop a thoughtful, well-rounded approach to fraud detection and prevention while maintaining a high-quality customer experience.