AI: The new teller

How can you tell when a new technology has emerged in the mainstream? One way is to look at how financial institutions are approaching it.

Companies that handle people’s savings, loans, and investments - not to mention sensitive data - face hefty demands regarding risk and regulations. These businesses must know their data and analytics are accurate, secure, and well-managed to be confident they are making the right decisions. At the same time, they need to prove to regulators that their operations meet stringent standards in every respect.

As a result, when a technology like Generative AI (GenAI) steps into the spotlight, financial services firms are typically more hesitant to adopt it than other industries.  When risk prevention and security are your top priority, you must manage every step of your transformation journey appropriately, with no data leaks or privacy concerns.

Change is underway, though. Financial firms’ initial hesitation is gradually being replaced by curiosity about what GenAI can help them achieve.

The tone of the conversation has also changed from mild curiosity to an enthusiastic quest for knowledge.

“Last year, our discussions were much more about information and knowledge sharing as well as thinking about what GenAI can do in terms of use cases and proof of concept (POC) projects,” adds Sharma. “This year, clients are asking us how to scale those use cases and POCs. They want to demonstrate the value GenAI can deliver so they can replicate it in other use cases, too.”

Fractal combines its expertise in designing innovative AI solutions with a deep understanding of the risk and governance issues involved. 

“Each business is unique in terms of its needs, investment appetite, and top-down support for technology innovation,” says Sharma. “Some are still watching and waiting so they can gauge the best direction to take. Others are jumping into GenAI with POCs. Use cases around employee productivity have been extremely popular, for instance, and these often involve using internal chatbots and virtual assistants. The organization can then use feedback from internal users to assess those use cases, figure out how to scale them up, and look at ways to open them to external users.”

Designing solutions for people, with people, is a core design principle at Fractal. This principle involves approaching each use case from the client’s unique perspective.  That includes considering the organization’s current position and needs and how it would evaluate, scale, and extend the solution to other users.   

Risk, complexity, and governance are vital to this approach for financial firms. Initially, these institutions are much more likely to embrace a use case with lower complexity and reputational risk.  When more complexity or risk enters the picture, they must know they have the structure and governance to move forward. 

"Now, integrating this cutting-edge technology with our well-established infrastructure demands a strategic hand. We’ll navigate this by adopting a phased approach, prioritizing areas where GenAI delivers the most significant impact. APIs and cloud migration will be our secret weapons for seamless integration.

But our commitment goes beyond efficiency. We understand the criticality of responsible AI.  We’ll be relentless in ensuring fairness by implementing robust data governance frameworks. We’ll invest in explainable AI tools to ensure transparency in decision-making. And most importantly, human oversight will remain a cornerstone. It’s about harnessing the power of GenAI while fostering trust and ethical considerations. 

At HSBC, we believe GenAI is a powerful tool with the potential to redefine financial services. We’re committed to wielding it responsibly for the benefit of our customers and the future of our business.” 

Building on these considerations, Sharma and his team have created a framework to help financial organizations plot their GenAI path. Its five broad industry-focused use cases, detailed below, provide a place to start.

Employee productivity is where many financial firms begin their GenAI journey, comprising three key areas.

Knowledge management engines allow employees to type in a query about anything from organizational policies to compliance and get a concise, accurate response. Instead of raising individual tickets for human resources, finance, or travel queries, users can have a fast, conversational experience based on information across the company’s data.  Based on user feedback, these internal-facing knowledge management engines can be rolled out to face sales agents and contact centers.  

Summarization engines can provide a similar, conversational experience tailored to roles like credit or analysts. These users create many different reports, so a tool that can summarize information in the relevant templates will streamline the process and save them vast amounts of time.

In the future, self-serve decisioning engines could also speed up financial decision-making.  However, this longer-term use case depends on a rock-solid structure, model validations, and governance measures.  

Customer experience is about creating more humanized, contextual interactions that develop and inspire trust—a crucial element for any financial services customer. That means finding ways to improve search functionality and dynamic content generation across the web and app while ensuring that every piece of information is complete and accurate.

Cost management and efficiency are the third use case, and this time, the focus is on the middle office and back office of the financial services. Fractal’s Senseforth solution is the star player here. Its capabilities, like chat summarization and real-time interaction insights, can be tailored to end-to-end processes such as underwriting to provide staff with a deep, contextual experience.

Fraud, risk, and compliance remain top concerns for financial firms, and GenAI can help address this two-sided coin.  

On one side, modeling the risk of fraud is an incredibly complex process. It can be even more complicated when the fraud is a high-severity, low-incidence event because there may not be enough data to build models that will identify it accurately. Synthetic data generation is one way that GenAI can help. It involves building enhanced data to stimulate existing information, enabling better modeling and identifying risk or fraud.

Making that information usable is the second side of the coin. How can the organization bubble the information efficiently and quickly to inform users’ decisions? The answer is to create a real-time, on-demand search and information platform for fraud and risk monitoring teams to engage with. This will give them the contextualized information they need to make the right decisions on the spot.

Fractal’s fifth use case, technology innovation, brings GenAI a broader engineering perspective in financial services. Its purpose is to find ways to help banks along their technology transformation journey. That might address data and cloud modernization issues, for example, as banks need to transfer code from legacy analytics platforms to new, cloud-based ones. Then there’s also the question of how to generate, translate, manage, and optimize those codes so they can be maintained consistently, securely, and compliantly. Fractal has developed accelerators that make it easier for financial firms to do that.

Considering these pathways is helping financial organizations to make significant strides on their GenAI journey. As organizations start implementing their ideas, the industry is experiencing an exciting time.  

“I evaluate any kind of a change in terms of three legs: acknowledge, assimilate, and adopt,” says Sharma. “Our clients in the financial services industry were originally apprehensive about GenAI, but they are now acknowledging that it is here to stay. They are assimilating the change. Most organizations in the industry have started to identify potential use cases to pursue, and many have found 20 or more. Momentum will increase now as more organizations seek to understand what their GenAI use case will be.

“This year’s big question is whether the assimilation we’re seeing today leads to adoption that unlocks value for those institutions. It’s very possible that the value unlocked in one use case can be replicated in another with great confidence in the results. It’s time for financial firms to explore those impacts and visualize the incremental value that GenAI can bring.”

Get AI delivered to your inbox!

Click here to subscribe to our quarterly magazine and stay ahead of the curve with exclusive insights, expert analysis, and the latest trends