AI Use Cases in Consumer Banking

1. Decode COBOL and PL/1 programs

In developed countries, many large banks, credit unions, and cooperatives still rely on COBOL or PL/1 software. These programs face several significant challenges: (a) these languages are nearly obsolete and not taught in schools, (b) most of the original developers have retired, (c) the code is often tangled and non-modular, (d) there is usually little or no documentation, and (f) the underlying code can be very hard to access. Since over two-thirds of the instructions in these languages are related to input and output, AI systems are being used to decode the entire COBOL and PL/1 code by training on the data available in the input and output tables. Trained AI systems have achieved over 98.5% accuracy in decoding these programs using only input and output tables, reducing manual labor and time by 75%. This also prevents mistakes in COBOL or PL/1 programs from being carried over when converting to a modern IT system. See www.scryai.com/data-flow-mapping for details.

2. Analyze transactions to detect fraudulent customer behavior

Use AI to detect fraudulent customer behavior in various financial products and services, including credit cards, Debit ACH transactions, wire transactions, phishing, and customer impersonation. AI systems can analyze transaction patterns and flag suspicious activities, helping to prevent financial losses and protect customers.

3. Analyze employee activities to detect fraudulent behavior

Use AI to detect fraudulent bank employee behavior, such as employees who have gone rogue or have conflicts of interest (e.g., investment bankers working for both the buy side and the sell side of a Merger-and-Acquisition transaction). This includes potential surveillance of employees’ social media and email communications. AI systems can monitor and analyze employee activities to identify risky behaviors and prevent internal fraud.

4. Improve compliance processes

Banks and financial institutions are regulated by governments, and these regulations change over time. Ensuring compliance is a huge task, so many banks use AI to improve regulatory compliance and workflows. AI systems scan legal and regulatory texts for compliance issues, handling thousands of documents without human interaction. GPTs and LLMs significantly improve these compliance processes. By automating the review of regulatory documents, banks can ensure they meet all legal requirements and avoid costly penalties.

5. Monitor personal finances

AI finance assistants help users monitor personal finances, such as setting target savings or spending rates based on users’ goals. These assistants can provide personalized financial advice and reminders, helping users stay on track with their financial plans.

6. Predict future bank balances of customers

Predicting customers’ future bank balances using AI systems helps users avoid overdraft fees and maintain their credit scores. Banks offer short-term loans to frequent overdraft users to help them avoid fees. By forecasting account balances, AI systems can alert users to potential overdrafts and suggest financial solutions.

7. Manage liquidity and plan for future deposit trends

AI systems predict deposit behavior and forecasts, including the likelihood of renewing term/fixed deposits. By analyzing historical data and customer behavior, AI can help banks manage their liquidity and plan for future deposit trends.

8. Identify non-performing loans

AI-based predictive models help identify consumer loans that are likely to be non-performing (i.e., not paid back). This helps banks discuss repayment strategies with customers and improve future credit applications. By identifying at-risk loans early, banks can take proactive measures to mitigate losses and support customers in managing their debt.

9. Improve the debt collection process

AI systems ensure a more compliant and efficient debt collection process. Borrowers often do not pay due to unwillingness or inability. AI systems trained on historical data understand these issues and help borrowers, reducing non-performing loans and assets. By tailoring debt collection strategies to individual circumstances, AI can improve recovery rates and maintain positive customer relationships.

10. Improve Customer Life-Time Value

AI techniques improve Customer Life-Time Value (CLTV) by identifying the next best product to sell, the best channel and time to contact the customer, and the right language and tone to use. By personalizing interactions and offers, AI helps increase customer satisfaction and loyalty, ultimately boosting revenue.

11. Understand customers' needs and preferences

Financial institutions use AI techniques to sell to customers at the right place and time, based on their current life events (e.g., marriage, birth, buying or selling a house, buying a car or appliance, or going on vacation). Customer behavior, interests, and hobbies (including social profiles) are also used to train AI systems. These AI systems help banks and credit unions gain a greater share of customers’ wallets and increase customer satisfaction. By understanding customers’ needs and preferences, AI can deliver personalized offers that resonate with customers, leading to higher conversion rates.

12. Frequency capping of messages

AI is used for “frequency capping of messages,” preventing customers from being bombarded with repeated advertisements, especially after they have already bought the item (e.g., a car). By controlling the frequency of messages, AI ensures a better customer experience and avoids annoying customers with redundant ads.

13. Determine the price elasticity of banking products

By analyzing past customer behavior and considering a cohort of similar customers, AI systems can quickly and accurately determine the price elasticity of banking products. This helps banks optimize pricing strategies to maximize revenue and customer satisfaction.

14. Optimize transaction denials

Most bank transactions that are denied currently use a fixed set of rules, which often do not work (e.g., if a customer has traveled outside the country). AI systems use 360-degree customer data to optimize transaction denials, e.g., if a customer bought a ticket from the U.S. to Germany and is using her credit card in Germany, that transaction is likely genuine. By considering the full context of a customer’s activities, AI can reduce false positives and improve the accuracy of transaction approvals.

15. Predict potential delinquencies

Decision Support Systems (DSS) predict potential delinquencies based on customer behavior and macro and microeconomic conditions. They also help lenders adjust credit lines appropriately. By providing insights into credit risk, DSS enable lenders to make informed decisions and manage their portfolios more effectively.

16. Debt recovery and customer support

Once a delinquency has occurred (or is about to occur), AI systems are used to understand and predict customer behavior, particularly their ability to pay versus willingness to pay. By distinguishing between these factors, AI can help banks develop appropriate strategies for debt recovery and customer support.

17. Match borrowers to mortgages

Using marketplaces and crowdsourcing, AI systems match borrowers to mortgage or auto loan lenders by pattern matching customer behavior with lender criteria. This approach helps borrowers find suitable loan options and lenders identify potential clients more efficiently.

18. Optimize administrative tasks

Using Computer Vision and Natural Language Processing, automated form filling is becoming common. Converting paper documents to electronic format using Intelligent Document Processing (IDP) is also widespread. These technologies streamline administrative tasks, reduce errors, and improve efficiency in document management.

19. Enhance marketing strategies

For credit cards, mortgage loans, auto loans, personal loans, and other banking products, AI systems are used for (a) creating better prospecting lists by including external data, (b) segmenting customers according to sales/marketing channels, (c) evaluating and improving campaigns, (d) helping customer representatives with the “next best action” while answering customer queries, (e) using attributes and predictors for improved credit underwriting, (f) using attributes and predictors for behaviors of good and bad customers, (g) real-time offers, e.g., if customers recently bought a house, they are likely to buy furniture, and (h) predicting total revenue and profit for that product segment (e.g., credit card or loans). Different banking products have distinct attributes, features, and predictors, leading to various use cases. By leveraging AI, banks can enhance their marketing strategies, improve customer service, and optimize credit risk management.

20. Profitability Forecasting

21. Optimize the performance of bank branches

Decision support systems analyze the performance of bank branches and ATMs, identifying areas needing improvement. By incorporating customer visit data, these systems help optimize the number of branches, their operating hours, locations, and staffing levels, ensuring efficient resource allocation.

22. Optimize the performance of bank branches

AI can predict which employees, particularly high-performing salespeople, VPs, and managing directors, are at risk of leaving. Similarly, AI can calculate the customer lifetime value (CLTV) and identify customers who might leave the financial institution, enabling proactive retention strategies.

23. Recognize employees who are likely to resign

AI analyzes customer behavior to determine the best products to market and sell to each customer at the optimal time, enhancing personalized marketing efforts and increasing sales effectiveness.

24. Enhance personalized marketing

AI optimizes customers’ financial portfolios by analyzing and balancing assets like stocks, bonds, cash, and real estate. It also optimizes retirement savings plans (e.g., IRA, Roth IRA, 401(K)) and provides personalized recommendations on what financial assets to buy or sell, ensuring better financial outcomes.

25. Optimize customer financial portfolios

AI automates the process of filling out forms and applications for various financial services, such as opening bank accounts, retirement accounts, wealth management accounts, and wire transfers, reducing manual effort and errors.

26. Automate form filling

AI categorizes customer complaints, both written and verbal, into different types. It then predicts the best course of action to resolve these complaints by analyzing past complaint resolutions, improving customer service efficiency.

27. Categorize and resolve customer complaints

AI systems strive to ensure fairness and non-discrimination when making offers or granting credit to customers. This is challenging because the concepts of fairness and non-discrimination are complex and often not fully understood even by humans.

28. Ensure fairness and non-discrimination

AI predicts cybercrime activities, including unauthorized access to customer accounts, credit card fraud, and phishing attempts, enabling proactive security measures to protect customers.

29. Proactive security measures

Customers occasionally face unique issues that existing AI chatbots or knowledge libraries cannot address, such as resolving fraudulent transactions. By training Large Language Models (LLMs) with internal bank data, banks can deploy Gen AI-powered chatbots that offer enhanced customer service experiences, speeding up processes like credit card fraud resolution.

30. Enhanced customer service experiences

AI matches financial clients with the nearest relevant information point or partner, such as a store for a promotion or an event like a product launch or fundraiser. It also provides customized reports tailored to customer requirements and habits, enhancing personalized service.

31. Match clients with the right information point

AI processes mortgages, loans, and credit requests, automating the credit approval process using advanced credit scoring techniques. It leverages various data points, including credit history, social media activity, geolocations, and browsing habits, to identify borrowers’ likelihood of repayment, improving decision accuracy.

32. Automate credit approval

AI supports decision-making on overdrafts, imposing consequences such as automated charges. It collects necessary information, aids in decision-making processes, and generates and sends letters and documentation regarding the outcomes of approval processes, streamlining operations.

33. Decision-making on overdrafts

AI manages debt collection and monitors the risk associated with it by analyzing economic, social, and contextual data. For instance, a drought could impact agricultural businesses’ ability to repay debts, or a decline in property prices could affect property agents’ debt repayment capabilities, allowing for proactive risk management.

34. Manage debt collection and monitor associated risks

AI identifies common patterns and early signs of customers at risk of credit default. It enables early interventions to reduce risk, such as requesting additional guarantees or offering extended credit periods. AI also adjusts credit limits and automates credit balance refunds, mitigating potential losses.

35. Predict credit default

AI manages, identifies, and assesses fraud risks, including fraudulent credit card use, web transactions, bank transfers, or checks. It analyzes historical transaction data, customer behavior, and demographics to identify common fraud patterns. When suspicious activity is detected, it alerts the fraud management team for further investigation.

36. Assess fraud risks

AI suspends credit cards or accounts when fraud is predicted. It manages communication with clients and insurance firms and uses geospatial analytics to identify unexpected and unusual transaction locations, enhancing fraud prevention measures.

37. Enhance fraud prevention

AI collects, indexes, and archives fraud evidence, including forms, police reports, and other documents. It supports fraud remediation workflows, ensuring that all necessary information is organized and accessible for resolving fraud cases.

38. Fraud remediation and resolution

AI manages the network of branches and ATMs, optimizing their locations using geospatial analytics based on usage data. It supports network and capital planning strategies by analyzing historical usage data, competition networks, and demand forecasts, ensuring efficient resource allocation and strategic planning.

39. Optimize resource allocation

AI manages the network of branches and ATMs, optimizing their locations using geospatial analytics based on usage data. It supports network and capital planning strategies by analyzing historical usage data, competition networks, and demand forecasts, ensuring efficient resource allocation and strategic planning.