The use of artificial intelligence (AI) in financial services is growing as AI offers multi-contextual benefits, including cost reductions, improved staff productivity, enhanced customer satisfaction, and reduced staff workload (Staista, 2021).
Given the way AI works and its operational model, it is well suited and is touted to help in the detection and reduction of financial crimes. AI enables the analysis of a large number of transactions and flags suspicious transactions for further analysis and investigation by the relevant authorities.
The use of machine learning technologies is considered an important development in AI-driven fraud protection efforts. The data fed into the system can help the system learn to identify anomalies and hence detect fraudulent and unauthorized activities. AI can help in making sound credit decisions and streamlining processes that lead to various financial approvals.
AI can help with fraud detection and mitigation across a wide range of industries, including banking, finance, insurance, public sector/government, and other industries (e.g., technology and service sectors).
According to a Statista report, improvement in fraud detection and cybersecurity are the key advantages of using AI. The report suggests that by combining supervised and unsupervised learning, organizations can get a better understanding of the behaviors and patterns of customers and users of systems, which can then enable them to identify fraudulent and suspicious transactions (Mlitz, 2021).
Crimes such as identity theft are on the rise, and customers lost an estimated $56 billion due to identity theft and fraudulent transactions in 2020 (Xie, 2021). Banks and financial institutions do perform checks to verify personal identities before approving loans. But with increased digital activity, people wanting to commit financial crimes are becoming smarter and trying to stay ahead of checks and balances to evade being caught during system checks. That is where AI can be very helpful in detecting anomalies and patterns to reduce the chances of such fraud. AI can help identify whether a person is using a stolen identity or a synthetic identity and mitigate against the occurrence of a financial crime.
The use of AI allows organizations to crunch multiple sources of data, including customer purchasing data, online activity data, credit bureau data, and personally identifiable information, to detect suspicious financial activities and take prompt action to stop a financial crime(Xie, 2021).As such, AI provides leverage to sniff out activities that otherwise can be very difficult to identify. Another advantage of using AI is that it can enable organizations to perform checks on transactional activities in real time, which can help counter fraudulent activities on an ongoing basis and mitigate losses to customers and organizations themselves.
What AI can or should do for fraud protection
Several things can be done to reduce the number of fraud cases each financial institution has. But, the first thing they must do is to start from the beginning by knowing each customer’s identity (biometrics and data verification with local authorities and international blacklists). Having the proper processes in place, we can eliminate identity theft, which is one of the biggest frauds globally.
Then, we need to know our customers’ behaviors. Financial institutions may have several types of personas that may be conducting fraud:
1. The cheater: Fraud against Financial Institutions.
This kind of customer saturates their line of credit quickly and fractionally (e.g., credit cards) without having recorded any payment towards the amount owed. They may also make payments that are less than the minimum allowed payment. So, the client has entered into “default” and it is required to focus the efforts on the collection process. AI can detect these types of behaviors. It can also monitor all the transactions that show probable unusual operations that indicate that the customer could be laundering money and/or engaging in illegal activities.
2. The cheated: Fraud against the customer
Identity theft from the account allows third parties to use customers’ accounts without the client’s having authorised it. This can be detected by verifying unusual transactions and comparing them with historical behavior, like the places they usually buy their groceries, have lunch, or have fun, and knowing when they are traveling. AI can help us identify this kind of behaviour to stop any transaction that may be illegal or not commonly done by the customer. As an example, a customer can’t be buying with his/her physical credit card at the same time at two places that are hours away from each other. Another example is when the customer is not used to buying from one specific online store. Here, we can detect this transaction and ask for a confirmation from the customer by sending him/her a message on their smartphone to approve that transaction (with a 2-step verification).
3. Operational errors: Mistakes done by the bank and/or the customer
Due to some type of human error, there may be abnormal transactions within the historical behavior of the client that will generate some kind of clarification, dispute, or some type of undue payment or collection. This will help financial institutions to provide better services to their customer base. AI may also detect these kinds of errors.
In the end, AI will help financial institutions to:
- Satisfy customers’ expectations (24x7x365) by putting the client at the center and giving continuity in the dialogue throughout the life cycle with the institution, massively personalizing the products and services, and making cognitive their institutional processes.
- Reduce operating costs and fines by increasing their efficiency (transactional and analytical) with an AI approach to their operational transformation, which also includes the risk and compliance processes. These also include a time reduction in the claims customers have, the number of fraud cases and anti-money laundering false positives, and the general cost of handling alerts.
Understand the causes of delay, the situation/solvency of their customer, their attitude, and their ability to pay to streamline the collection process to increase its effectiveness from delinquent customers.
Professor Jiwat Ram
CIO Consultant – Advisor | Digital Transformation Director
Latam Business School · Part-time
Mlitz, K. (2021). Global AI use cases for financial services 2020. https://www.statista.com/statistics/1197955/ai-financial-services-global/
Statista, (2021), Main benefits expected by financial services firms from adoption and use of artificial intelligence (AI) worldwide in 2020, by region, https://www.statista.com/statistics/1246927/financial-services-benefits-from-ai-adoption-2020/
Xie, Y. (2021). Is Identity Theft Fraud? Here’s How AI And Graphing Tools Help Solve The Puzzle. https://www.forbes.com/sites/forbestechcouncil/2021/09/30/is-identity-theft-fraud-heres-how-ai-and-graphing-tools-help-solve-the-puzzle/?sh=77b108787845