What AI is and how it works in the financial services industry—Part 3

29 SEPTEMBER 2016
By Parth Desai

Parth Desai, Founder and CEO and AI guru discusses the results of the recent Finextra survey on the appetite and adoption of AI in Financial Services and asks, "What are the odds of AI based solutions becoming the imperative—not a nice to have?"

If I were a betting man, I would have placed a large bet on there being a different outcome from the recent Finextra survey on the adoption of AI in financial services. To a person within Pelican, we all predicted that compliance and security would be top of everyone's agenda. In fact this was not the case. Although it still figures in the list of priorities, it was time-to-market and product innovation that now occupy the top spots. This was a surprise at first, but having had time to think about the results, I can now see why this would be a major concern and I will talk more about this later.

Next on the list, which was expected, is the continuing headache of dealing with legacy systems. Within payments processing, repair, routing and investigations were also found to be inefficient. And one can almost sense the frustration many banks are feeling because, despite significant investments in payment processing systems and compliance, they still remain highly inefficient.

So looking at time-to-market and product innovation, both of which were cited to be among the most inefficient areas in banking, but are crucial to the future success of the business. This could be because of several reasons, including payments themselves which are not changing. It's actually the usage, integration and user interface demands that are dramatically changing. This requires a completely new approach incorporating rich and intelligent interface technologies like natural language processing (NLP) and omni-channel UI.

So what's next—the $64,000 question has to be—can new technology resolve some of these longstanding issues? The results of the survey show that a significant number of respondents are clearly on the lookout for new solutions—especially within product innovation and reduced time to market. Sanctions, AML systems, payment repair and exceptions processing systems are also appearing on their shopping lists. 

We were very pleasantly surprised to learn that 70% of the respondents believed that the adoption of AI will increase significantly in the payments industry over the next two years. There was also a strong agreement [45%] that AI has the potential to address the remaining inefficiencies in the payments business. However, there are still a number of adoption hurdles to overcome, as some respondents think that they would struggle to get management buy-in. Although in contrast, a sizeable number [38%] believe that AI can be relied upon and some [11%] are already using AI in their payments processing.

AI and Banking
We know that Artificial Intelligence is solving several challenges faced by the payments industry and also reducing many of the inefficiencies. Our experience in AI spans over 20 years, and we are already supporting long standing customers in areas such as sanctions processing, payments repair, exceptions and intelligent least-cost routing. Those customers who are using AI based solutions have been able to reduce, or virtually eliminate, the high levels of human intervention and manual processing that were previously necessary. This has enabled them achieve lower costs, accelerate processing time and reduce errors across the board. Other areas currently being explored with our customers are automation of exceptions & investigations and customer retention.

One area where banks will derive significant benefits from AI is when they start to focus their attention on product Innovation. By exploiting the vast and invaluable  amounts of data that banks possess on customers behaviours and preferences, via technologies of machine learning, valuable insights will be acquired enabling new, and relevant, products and services to be created. Use of knowledge based approach in combination with the right AI engines, time to market can also be significantly reduced. This will help banks to compete more effectively against the new fintech players and challenger banks. In future, we also feel the use of voice recognition combined with natural language processing will be a powerful combination enabling more innovation and a significantly improved user experience.

So all in all, based on the feedback from the Finextra survey, it appears that the future of AI is pretty much assured. Indeed, if I were a betting man, this time I would be inclined to place a very large bet that within the next two to three years AI based solutions will be a Business-As-Usual requirement. Not a nice to have.

Read Part 1 and Part 2 of this series here.

Back to Articles