GenAI Use Cases: Banking
Financial intermediaries connect savers and borrowers by facilitating the flow of funds through institutions like banks, credit unions, and investment firms. These intermediaries manage risk, provide liquidity, and allocate capital efficiently across the economy through services such as lending, deposit-taking, and asset management.
Key players: Commercial Banks (BofA, JP Morgan Chase, etc.), Investment Banks (Goldman Sachs, Morgan Stanley), Credit Unions and Savings Instiutions (Navy Federal Credit Union, Ally Bank), Government Sponsored Enterprises (Frannie Mae, Freddie Mac), and FinTech Platforms (SoFi).
The McKinsey Global Institute estimates that across all of banking, wholesale, and retail, GenAI could add between $200 and $340 billion in value. Not only is GenAI capable of creating content for drafts, but it is also adept at understanding previously published content. New product development, customer operations, and marketing and sales provide the largest potential for the new technology.
New product deveopment: JPMorgan Chase has filed a patent application for a GenAI service that can help investors select equities. Morgan Stanley has built a tool to help RMs deliver relevant ideas to customers in real time. Banks are also using GenAI to speed code development.
Customer operations: Banks are using GenAI to extract, search, and summarize unstructured servicing information and translate it into machine-readable instructions. It’s also automating manual tasks as it is proving capable of writing technical documents such as financial; environmental, social, and governance (ESG); and audit reports. It’s also being used to write loan contracts such as mortgages.
Marketing and Sales: GenAI has the potential to take all voice and text interactions with clients (and internal discussions about the client) and use them to create an “Relationship Manager assistant.” An RM assistant can help with tasks such as investment ideas, sales, and product policies nearly instantly. GenAI can help junior RMs better meet client needs through training simulations and personalized coaching suggestions based on call transcripts.
According to McKinsey, a leading investment bank has built a GenAI tool to help analysts write first drafts of pitch books. The analyst uploads all the relevant documents and then instructs the chat bot to produce many of the slides that are typically needed and many others that reflect the specifics of the proposed investment. “The tool saves analysts about 30 percent of the time they used to spend creating pitchbooks,” the report states.
Banks are cautiously expanding into generative AI to further enhance both customer-facing and internal functions. One prominent use is in customer communications: banks are using large language models to draft personalized emails or explanations for customers, for instance when denying a loan (to ensure clarity and compliance in the reason given), or to summarize new product offerings in plain language. Banks must be careful to meet regulatory requirements (fair lending laws, consumer protection) when using AI in this manner, so many are adopting a “human-in-the-loop” approach. Another growing area is using generative AI internally to support employees. JP Morgan and other big banks, while initially banning staff from using public ChatGPT for confidentiality reasons, have been developing proprietary large language models trained on internal documents to assist workers (Forbes). These AI helpers can instantly answer bankers’ or analysts’ queries from vast troves of manuals, procedures, or research reports. In credit underwriting, generative AI can scan and summarize borrowers’ submitted documents (tax returns, business plans), freeing officers to focus on decisions. Some banks are exploring AI to generate synthetic data for model testing, which could improve risk models without using sensitive real customer data. Importantly, regulators have increased their scrutiny of AI in banking. The Federal Reserve, OCC, and FDIC have issued model risk management guidance (e.g. SR 11-7) that requires banks to rigorously validate and audit their AI models for accuracy and bias. Banks also face potential rules around AI fairness – for example, ensuring that AI-driven credit decisions do not inadvertently discriminate against protected classes, which is a hot topic among policymakers. In summary, AI in banking has evolved from automating credit decisions and back-office tasks to powering intelligent customer interfaces and complex risk analysis. The integration is so complete that virtually every bank transaction today – from applying for a loan to a fraud check on a debit card swipe – likely involves some AI or algorithmic model making a judgment in the blink of an eye.
See the McKinsey graphic below for a representation of the full-suite of capabilities AI can bring to banking
