Real-time drives AI uptake

By Parth Desai | 20 December 2016

This has been a defining year for artificial intelligence (AI), with banks and corporates starting to realise its potential. Recent developments in big data and cloud computing are certainly aiding this uptake, but for me the key driver in the world of payments is the move to real-time. Real-time payments kicked off in Europe and the wave of change had travelled east around the world and is now in North America. This wave is continuing to roll, and will make a second lap around bringing with it a more agile approach.

Early adopters such as the UK can benefit from the lessons learned from more recent adoptions. In fact, the next innovations are already taking place such as ISO 20022 and real-time gross settlement.

The imperative to change

Opinions about the implications of real-time payments vary across geographies. While many people in countries where real-time payments are already in place show a more relaxed approach to the implementation, and the technology surrounding it, this feeling of confidence needs to be more widespread for real-time payments to work effectively.

The deployment of AI is an absolute necessity to support real-time payments. This means that everyone needs to take up the technology, regardless of country.

Imagine if, during your experience of using Uber, the normally frictionless payment was held up so your identity could be checked. It wouldn’t happen of course because the whole Uber model makes the payment process disappear. But this serves as a good illustration of the problems that could occur if, when the real-time phenomenon makes itself felt beyond retail payments in the corporate and transaction banking environments, all the other activities associated with those payments fail to keep pace.

Put bluntly, with the move from batch and multi-day to real-time in transaction banking, there won’t be any time for human intervention in payments. We are all familiar with real-time for cards, with Faster Payments in the UK and similar schemes elsewhere. The move up the value chain to corporate payments is still to come. But it is not far away.

In fact, the first use cases for real-time payments in the US are business payments. The revenue potential of real-time is almost entirely driven by high-value payments, and that is increasingly where the focus will be. In addition, I believe real-time will go cross-border within five years.

To underpin real-time payments, all the supporting processes – AML checking and various customer service-related activities – will need to happen in real-time as well, and this is where AI can play a major role.

Introducing AI

The situation in payments today is unsustainable. Many banks are running significant operations departments full of people processing payment exceptions that don’t meet the criteria for automation. A solution is required to slim down this back office function.

AI can bring automation to exception processing, handling investigations, fraud and sanctions checking in real-time. This will allow banks to deliver on real-time with the necessary efficiency and speed.

The good news is that AI is not an extraordinary technology. AI is simply logic, gathered via human input or machine learning, coupled with very fast computers which allow that logic to perform very quickly. Indeed, AI now makes sense as a technology that can be realistically implemented to carry out functions at scale and speed, and is ready for mainstream use to support the widespread roll-out of real-time payments.

Banks already using AI-based solutions have been able to reduce, or virtually eliminate, the high levels of human intervention and manual processing previously necessary. This has increased efficiency, accelerated processing times and reduced errors across the board as well as achieved significant cost savings.

Looking to the future

It is important not to view AI in isolation. We are on a trajectory in payments, moving from the old world to the new.

No bank will make this happen in a very short period of time, but each must understand its own product roadmap for the next five to ten years. The firm’s response to the real-time transformation of transaction banking, open API-based initiatives like PSD2 and then map the platforms it will need to deliver them, will help them to create a competitive differentiator.

The areas in which AI should be brought to bear will become clear, not just to solve problems today, but to enable future payment opportunities tomorrow. The industry needs to gain a better understanding of how AI technologies apply in open and real-time transaction banking, to move away from a situation where people in the business who understand the current operations don’t understand what the technology can do, and the technologists have no idea about operations or the complexity of payments.

I believe AI technology is here to stay, and uptake will increase in 2017 as it gradually becomes part of the mainstream. Early adopters of the technology will gain a competitive advantage – helping them survive and thrive in an open real-time world.

Parth Desai, Founder and CEO, Pelican

Become a bobsguide member to access the following

1. Unrestricted access to bobsguide
2. Send a proposal request
3. Insights delivered daily to your inbox
4. Career development