The rapid evolution of artificial intelligence (AI) and its subset, Generative AI (GenAI), are creating a world of new opportunities in the banking sector. These were discussed during HCLTech’s latest Data and AI Forum for Banking in London.
“Recently, there has been a huge focus on data and AI, and generative AI specifically, in banking. Over the last few years, we've made significant strides in setting up AI and FinTech labs, including one in London, to help push forward AI innovation across the banking sector,” said Santosh Kumar, Senior Vice President and Head of UK Banking Business at HCLTech as he welcomed attendees, highlighting the profound impact of AI on customer satisfaction and operational efficiency in the financial services sector.
During the event, Kumar was joined in a fireside chat by Paul Weller, Former Head of New Customer Propositions & Conversational AI at NatWest, and they explored the evolution of customer-facing chatbots, the importance of data quality and governance, regulatory challenges and a new era of fraud prevention.
Cora: A journey from conversation to generation
Cora, NatWest’s conversational chatbot, has undergone a significant transformation since its launch in 2018 as a question-and-answer bot. It was originally developed using IBM Watson and in July 2024, integrated WatsonX Generative AI components.
Sharing insights into Cora's journey, Weller said, “Some customers were initially hesitant about using bots, especially certain demographics who prefer speaking with humans. However, over time engagement has increased and some groups now prefer speaking with our automated solution. Last year, Cora handled around 16 million conversations across our mobile app, website and call center integrations.”
He added, “If we were to remove Cora, we would need a substantial number of additional human agents to manage those interactions. It’s a massive saving for the organization and a benefit for customers, as it provides quicker and easier access to answers.”
One unexpected benefit Weller highlighted was during bereavement processes. NatWest found a majority of customers prefer interacting with a bot rather than speaking with a specialist bereavement handler. Similarly, neurodiverse customers have found Cora’s succinct responses and intuitive design more accessible.
The introduction of GenAI has further streamlined customer interactions. The AI can now respond to complex inquiries — such as explaining mortgage terms while providing links to the relevant materials — efficiently. Customers can opt into using GenAI, and an impressive number choose to do so, with high satisfaction ratings.
Data quality and governance: The backbone of AI success
Implementing AI in banking is not without challenges. Data quality is paramount.
For banks with decades of legacy customer data, maintaining accurate, consistent and actionable datasets is critical. Contradictory data — such as outdated HR policies or inconsistent customer records — can lead to unreliable AI outputs. Addressing this requires significant investment in data governance frameworks. Banks must ensure their data is updated, accurate and ethical in its use.
“If the information that you’re feeding to the LLM or SLM behind your bots is flawed, the information that comes out will be flawed,” confirmed Weller.
He also stressed the importance of clear organizational ethical practices. “AI has no ethics. It adopts the ethics of the organization it serves.”
This underlines the need for robust oversight, ensuring AI aligns with a bank’s customer-centric values while mitigating risks such as bias and misrepresentation.
Navigating regulatory hurdles
Regulations pose another layer of complexity. The financial sector is highly regulated, with new AI-specific guidelines emerging globally. The European Union, for example, has introduced detailed AI regulations, which are likely to influence UK policies.
To overcome regulatory challenges, banks should start with low-risk use cases like customer FAQs and then gradually build trust with regulators and customers alike. Engaging with regulators early, especially for high-stakes areas like financial advice, is critical.
Fraud prevention: Combating emerging threats
AI’s potential extends beyond customer-facing solutions. One of the most critical internal applications is fraud detection.
Kumar shared a key example of AI’s role in improving trade surveillance for fraud detection.
“One of the large investment banks that we work with adopted GenAI for trade surveillance. It processes thousands of alerts and conversation data from emails or chats, prioritizes them and generates summaries for analysts. This reduced the manual workload and enhanced efficiency by 60%.” Such use cases highlight the importance of GenAI in tackling vast amounts of data to flag high-risk activities and support fraud detection.
Additionally, Kumar discussed another use case. “Different customers and countries have their own ways of doing documentation, with some still using manual methods like faxing. We’re exploring how we can use AI to read and summarize these documents for operations teams to improve efficiency. In this case, we’re leveraging Google Cloud’s Document AI, and we’ve already seen a 30% increase in efficiency.”
Weller added that AI empowers organizations by focusing resources on true pain points, especially in fraud detection.
“It gives them the ability to see today what they should be losing sleep over tonight. What they were seeing before was something that happened a month ago, and they were only just now losing sleep about it. AI and augmented tools really help teams get on top of what they need to do, shortening the time to act and improving their response.”
However, it should be noted that fraudsters and cybercriminals are increasingly leveraging AI to create hyper-realistic schemes, such as AI-generated deepfake videos.
Driving customer satisfaction and business efficiency
The banking industry's embrace of AI, particularly GenAI, marks a new era of innovation. Solutions like Cora demonstrate how AI can enhance customer interactions, save costs and address sensitive and complex issues quickly. However, success hinges on data quality, robust governance and collaboration with regulators.
Moving forward, embedding AI thoughtfully within processes is crucial. It’s not just about cost savings; it’s about efficiency, better decision-making and empowering teams with the right tools to address critical challenges. The adoption of AI and GenAI will continue to create value, enabling banks to serve customers better while achieving operational excellence.
