09 July 2024
Our Chief Revenue Officer Khilan Shah recently spoke on the ‘Leveraging LLMs to enhance regulatory data’ panel at this year’s Risk Live Europe event in London.
Alongside panellists Andreas Ita and Andrew Mackay, and moderator Zoi Fletcher, Khilan explored how Large Language Models (LLMs) can transform data management within investment banks, touching on the high barriers impeding their implementation and what needs to be done to overcome them.
We caught up with Khilan and Emily Martyr, Events Director at Risk.net, after the panel, to dive more deeply into what was discussed on the day.
Emily:
Risk Live Europe, hosted by Risk.net in London between 18-19 June, is a key event for financial risk professionals. This flagship European festival brings Risk.net’s expert editorial content to life on stage, providing crucial networking opportunities for risk management experts.
Khilan:
Risk.net is the de facto association for risk management and regulation in Financial Services. So, it made sense for Likezero to get involved at Risk Live and share our expertise on the transformative potential of AI for data management and compliance. The opportunity to exchange ideas with fellow experts was unique, as were the valuable insights gained from audience polling during the panel I spoke on (more on this later).
Khilan:
Being deeply involved in this field, it’s easy to slip into technical jargon that may be inaccessible to those not engaged daily. This panel was effective in simplifying the topic and highlighted that AI is still very new to much of the banking industry. Even with top talent, many are just beginning to understand what AI is and the potential it holds. This underscores the rapid pace of AI advancement. But, it’s crucial to step back and grasp its potential before moving forward. Banks and financial institutions clearly need a better understanding of how to effectively leverage AI, including where to start.
Emily:
The audience was definitely eager to learn about real-world AI applications and how these capabilities can be implemented in financial institutions. This is a significant challenge, and partners like Likezero are important as the industry seeks to safely leverage AI to enhance risk management practices.
Khilan:
It was also evident that AI has opened a Pandora’s box for financial institutions. When you talk to them about AI there’s a noticeable buzz, as they recognise their transformative potential. However, there’s also some nervousness about their implementation due to a lack of understanding of how they work and how to manage the data output. It’s essential to define data needs from the ground up and understand what the data represents and how to use it. The panel was excellent in encouraging the audience to think practically about using this information.
Khilan:
In the poll, the biggest challenge investment banks face with LLMs, 48% of the audience pointed to AI governance, and 38% mentioned data quality and consistency. This reflects the fact that AI governance and data quality are especially pertinent concerns for banks and institutions. While many do expect new AI technology to improve data quality, they’re also beginning to consider the complexities of implementing LLMs. Banks, being heavily regulated, face unique challenges with AI. They must ensure they understand and can rely on AI-generated outputs when regulators review them. The process involves some uncertainty, which is normal since every model is based on certain assumptions. But recognising this uncertainty is part of governance. Transparency is essential, too – banks must disclose potential issues and clearly explain the assumptions and risks involved. This openness is key to addressing concerns and ensuring compliance with AI technology.
Emily:
In my view, it’s not just about finding the best use cases for all businesses. In the financial services sector, implementing AI requires use cases to be accompanied by a level of control appropriate for highly regulated organisations. By collaborating with regulators, investing in innovation centres and sharing best practices with peers, we hope to see more validated applications that can benefit financial institutions and their clients.
Khilan:
The audience overwhelmingly identified automating data extraction (73%) as the top opportunity for leveraging LLMs in regulatory compliance and risk management, compared to 15% for enhancing risk models. This isn’t surprising given that automated data extraction is a common LLM application, reducing the manual effort of sifting through information. It’s a low-risk starting point, where LLMs can generate results for human verification to ensure accuracy, addressing some governance concerns. However, achieving better decisions hinges on automating data extraction – they’re two sides of the same coin. We need to start talking more about how LLMs enhance decision-making, rather than focusing primarily on its ability to automate data extraction.
Emily:
I think opportunities vary for each firm and specific teams based on their best business use cases. However, a common area of opportunity lies in efficiency gains, which save teams time and allow them to focus on high-value tasks instead of repetitive work that AI can handle. We heard a great overview of AI’s benefits from Deutsche Bank on a separate Risk Live panel.
Khilan:
Many assume only lawyers handle contracts, but they work for the business and don’t bear personal risks from contract terms. This assumption leads other team members to shy away from legal matters, labelling them as ‘contractual issues’. Without centralised data management, inconsistencies spread, complicating regulatory compliance and undermining risk management. Access to timely, accurate data is critical, yet without a central data source and consistent maintenance, this becomes a challenge. This is where Likezero comes in, offering a central source of data where all teams can access and gather uniform information. We provide our clients with a centralised repository of documents and data, providing quick insights and analysis, which helps manage risks and improve decision-making.
Emily:
I would also say that the importance of contracts became clear during the switch from LIBOR to other reference rates last year. I think senior professionals now realise that everyone needs to help manage contract information and data across departments. In today’s AI-driven world, having many sources of data is more valuable than ever, and companies know that good data management should be a regular part of their operations.
Emily:
I’m extremely excited about our next event. It’s going to be great – more themes will be covered, an exceptional line up of speakers and more networking opportunities than ever before. Plus, we’ll also be introducing an awards gala dinner so that we can celebrate industry innovators in style.
Khilan:
Me too. We’re looking forward to getting more involved next year. It presents another fantastic opportunity to drive these critical conversations forward.
