I think the journey to generative AI is really reshaping the banking space. We see, broadly, three categories of adoption, the first one being back-office operations, and the second involving the risk space. The third category is customer-facing and includes either a copilot supporting the frontline or direct interaction with customers.
On the back-office side, gen AI leverages large language models to automate workflows and reduce internal friction or the need for outsourcing, allowing processes to be done faster, cheaper, and far more accurately. In the risk space, gen AI is impacting the more traditional underwriting models, collection, and monitoring systems, again, making everything faster and more accurate.
On the customer-facing side, we’re talking about two areas. One involves what we call a copilot, which helps frontline relationship managers who interact directly with customers, be they corporate, business, or relationship banking on the retail side. Copilots don’t replace employees—they enable them to provide better service. So instead of a relationship manager reading tons of material to prepare for a customer discussion, the copilot leverages AI capabilities to synthesize all the client knowledge to make recommendations a timely manner. This enables better, more targeted discussions with clients.
In terms of direct interaction with customers, that’s something we see a lot more of in emerging markets, like Asia. On the European side, banks are focusing more on the back-office risk functionalities and exploring opportunities in the copiloting of customer interactions. The US is somewhere in between, depending on the size of the bank, and remains in the early stages of exploring different areas. I think we are only scratching the surface, as we learn more and get comfortable with the limitations of the technology—as well as the risks and regulations. We think everything will evolve faster in 2024.