All financial institutions are aware of the need to implement solutions that leverage the benefits of AI. But few use cases ever pass the experimental stages, according to Broadridge’s Joseph Lo in an interview with finews.asia.
The vision of leveraging artificial intelligence (AI) capabilities to improve businesses in the financial sector has become a ubiquitous matter. This is not just a matter of trendiness but also improving the bottom line. According to a Citi report, AI is expected to add $170 billion in profits for the banking industry by 2028.
Despite the ambitions, there is still a gap between visions and realization.
«Every financial institution we speak to says they know they have to use AI. There is downward pressure from boardrooms, particularly to reduce costs,» said Joseph Lo, head of enterprise platforms at Broadridge, in an interview with finews.asia. «However, there is a struggle to start.»
Organizational Challenges
According to Lo, many are still stuck in the experimental stages due to organizational challenges.
He observes that one trend that has driven success is to organize AI into a standalone area of focus, such as the creation of an AI center of excellence rather than being combined with general innovation and risk being under-prioritized. He also stressed the importance of balancing insourcing and outsourcing as well as using service providers as «platforms rather than a single-point solution».
«In all cases, we’ve seen successful pilots and proof-of-concept (POC). But it’s rare for AI projects to pass the POC phase,» Lo reiterated.
Top Use Cases
On leading use cases being explored or implemented, Lo highlights three areas: faster access to data, improving client experiences and enhancing productivity.
In terms of data access, Broadridge launched «BondGPT», the world’s first large language model-powered bond discovery tool which is helping 900 users quickly find related information like price or liquidity. Client experiences are being improved by using AI to create personalized content such as custom newsletters. Optical character recognition and natural language understanding capabilities are boosting productivity by extracting relevant data from documents.
High Risk Areas
But the use cases are not limitless. There are certain areas that Broadridge considers high risk and avoids delivering related solutions.
This includes tools that are completely automated with no human intervention or those that detect emotions. The firm also avoids tools that can potentially affect livelihoods in a negative manner in areas such as hiring decisions and performance management.
Geopolitics in AI
And there are some challenges that are especially applicable to the region, such as geopolitics.
«One of the unique risks in Asia is that many financial institutions are less comfortable with SaaS (software-as-a-service) solutions and are looking more at on-premise solutions,» Lo explained.
«A lot of AI technology was invented in the US, which is at the forefront of this industry. There is hesitancy to rely on such technology that can’t be influenced, especially due to the political uncertainty in the US. From our experience in the APAC region, there is significant focus on data sovereignty and concern about the risk of over-reliance on foreign providers.»
Asian Hub
Aside from users in Asia, there are also various business centers vying to be AI hubs by attracting both capital and talent. According to Lo, there is one particular standout within the region.
«In my opinion, Singapore is one of the leaders of AI in Asia,» he said. «I think this is because their government is very open to AI innovation and inviting world class talent to work in the country, all while balancing the need for safety, security, and privacy. Their population also has a strong preference for consumer-based innovation.»
