The rise of artificial intelligence (AI) is poised to revolutionise industries across the board, and banking is no exception. Reginald Warlop, an executive specialised in digital transformation, offers his insights on how generative AI, a particularly powerful branch of AI, will impact the future of banking, and the challenges and opportunities it presents.

A shift from specialised tools to widespread impact

Advancements in recent years mean that AI has transitioned from a niche tool for specific tasks to a more accessible technology with broader applications. Generative AI can now be applied to fields like customer engagement, marketing, and knowledge management, significantly impacting how banks operate internally, as well as how they interact with clients.

“Initially, it was really a very specialised activity, so resources were scarce and mainly deployed towards specialist algorithms like propensity models and retention modelling or fraud modelling,” explains Reginald. “Now with the advent of generative AI, it’s much more accessible. It’s going to increase the impact it has on operations, especially in the fields of customer and client engagement and communications, marketing, software coding, or just accessing knowledge bases or spreading knowledge internally within organisations.” 

While the potential is vast, Reginald acknowledges the challenges that come with generative AI. Models can exhibit unpredictable behaviour over time, requiring robust monitoring and risk management protocols. Additionally, data privacy and compliance regulations surrounding this evolving technology are still being established, adding another layer of complexity for users. 

“The challenges are multiple, for example you create a large language model that you’ve trained on your own content, it starts deviating over time, it starts behaving differently: it’s not like software coded once that always behaves exactly in the same way,” offers Reginald. “Because of that, there’s a number of consequences in terms of operations, model monitoring, but also risk compliance, security.

“There’s loads of different dimensions to that problem. A number of measures needs to be put in place to really use the technology at scale 24/7.” 

The dawn of conversational interfaces

One of the most exciting aspects of generative AI is its ability to transform the way we interact with technology. Reginald envisions a future where rigid, menu-driven interfaces are replaced by fluid, conversational experiences. Imagine asking your bank questions and receiving personalised responses tailored to your individual needs without ever speaking to another person. This shift represents a fundamental change in how people interact with banking services.

“What excites me is the fact that it changes the whole interface: with computing, traditionally applications and ERP systems, enterprise systems all have evolved in a similar way whereby you look at a set of screens with a number of fields, boxes, drop downs, a regimented hardcoded process,” highlights Reginald. “I think with generative AI, you can extract out all of this kind of box-based, tick-box based experience. It becomes a much more fluid conversation.

“The computer can ask you questions and in return the process will adapt. The process becomes much more flexible.” 

Generative AI’s ability to personalise experiences based on individual characteristics unlocks significant possibilities. Customers could begin to receive financial advice or product recommendations tailored to their unique financial situation and risk tolerance at a level that hasn’t been seen before. This degree of personalisation has the potential to significantly improve customer satisfaction and engagement, creating stronger relationships between banks and their customers. 

The importance of governance

While the decentralised nature of generative AI allows for experimentation and innovation, Reginald emphasises the importance of establishing governance frameworks. Organisations need to control the use of public models, set clear guidelines for development and deployment, and ensure compliance with evolving regulations that are yet to be released. Failing to do so could lead to security risks. 

“There are a number of challenges around the regulations which are not mature yet, and so what you’re doing now may have an impact on how you will comply to regulations that may come out in a year’s time,” emphasises Reginald. “Generative AI given that it’s so accessible, it’s very easy to get started. Loads of people are trying out something new and in some companies, you have hundreds of those little tests going on because anyone really can do it with a little python script.

“It’s very important that organisations contain that, set a governance framework around it to make sure that if employees are using public models, that’s being shut down ASAP, and to really put down a set of rules.” 

Generative AI holds immense potential to transform the banking industry. From personalised customer experiences to streamlined operations, the benefits could be endless. However, navigating the challenges related to data privacy, model governance, and evolving regulations will be crucial for banks to successfully harness this powerful technology. By approaching this technology with a cautious yet optimistic lens, banks can unlock a future where AI empowers them to deliver exceptional customer service, optimise operations, and stay ahead of the curve in an increasingly competitive landscape.

Keep up to date with the latest Banking Transformation Summit news on LinkedIn.