The Value of Embracing AI in Payments

Artificial intelligence (AI) technologies are well established in the banking industry across a range of workflows, customer-facing services, and risk and compliance initiatives. However, this space is evolving rapidly; recent developments in the technology are seeing a growth in investment in AI across a range of areas.

Payments, in particular, may benefit from the opportunity presented by AI, thanks to the richness of transaction data (particularly in ISO 20022 formats) and the complex nature of processing non-card payments. While using data in payment messages as a method to enhance services, including customer service (CX), has been understood for years, the industry is now moving toward additional ways that bank can optimize their usage of payments data. Banks can turn to modern data technologies for a range of use cases to improve operational efficiency, support margin growth by lowering costs, or further improve CX.

The Evolving Application of AI in Banking

Artificial intelligence, as explained by Celent’s working definition, “refers to the simulation of human intelligence in machines that are programmed to think and learn like humans. It involves the use of machine learning, natural language processing, data analytics, and other AI tools to analyze large volumes of data, make predictions, detect patterns, and provide personalized recommendations and solutions to customers.” AI technologies include computer vision, expert systems, machine learning, natural language processing, neural networks, and robotics.

Generative AI (Gen AI) is another AI technology. While not a new concept, recent developments in this area have opened up new thinking about the way the technology can be used to support a range of different workflows. Generative AI is already being used by some banks to support relationship managers and advisors. Yet more innovation is anticipated. Today, as found in the Celent Technology Insight and Strategy Survey 2023, with responses from a panel of corporate banking executives, nearly 3/5 (58%) of banks are already testing or evaluating Gen AI in some capacity, with an additional 23% having Gen AI projects in their current 2023/2024 roadmaps. Additionally, more than a third (36%) see this as a technology that will have the most significant market impact in the coming five years.

As banks establish their investment agendas, AI and advanced analytics are proving to deliver clear revenue opportunities. Now is the time for banks to focus on the potential outcome of those investments. Banks must not only determine and embrace the use cases for AI in payments; they must also consider implications for sensitive customer data, regulatory complexities, and the explainability and auditability of LLM outputs.

Emerging Use Cases for AI in Payments

Use cases for artificial intelligence in payments take various shapes, delivering benefits across the front, middle, and back offices. These can variously support improved efficiency and unlock new value for the bank and its customers.

Robust insights for the front office. Customer-facing services at financial institutions can benefit by the use of AI in multiple ways. First cash flow analysis and forecasting can use the information in payment messages to deliver granular, real-time insights into a corporate client’s cash positions. AI, used with payments data, can support use cases around liquidity and management to help corporate clients optimize their working capital. Additionally, Gen AI shows promise for providing customers with account insights—such as the ability to ask questions about payment status or to request a visualization of inbound payments over a particular timeframe—and (crucially) delivering accurate answers.

Workflow and process improvements for the middle and back office. AI technologies are already in use for a range of middle and back office functions related to payment processing, but can also offer additional improvements for customer service and/or cost reduction. Payment processing can be optimized by using ML to automate the payments repair process; smart automation can also increase straight-through processing (STP) rates for error reduction and repair. AI can improve transaction routing. These technologies can also help translate or convert messages, such as by translating unstructured names and addresses in ISO 8583 into the structured format required for ISO 20022 messages, a transition that otherwise may be “costly and potentially complex.” AI can also be applied to risk and fraud initiatives, detecting financial crime and fraud, while reducing the reliance on time-intensive, manual processes for transaction screening and the review of false positives.

Improved agility and efficiency. The concept of agility is becoming increasingly essential for financial institutions. Key to achieving that is the ability to meet market opportunities and customers’ rapidly changing needs. Currently, the lack of developer capacity is one of banks’ prime challenges when it comes to innovating around payment products. Opportunity costs related to the failure to deliver product enhancements are significant; while not entirely quantifiable, banks estimate that product enhancements that they couldn’t deliver due to resource constraints would have supported more than 5% growth in payments revenues over the past two years, as found in to the Celent Low Code in Payment Processing Survey 2023. Using Gen AI to support efficient code generation, optimize code, create documentation for code, convert code, process user feedback, and to increase developer efficiency overall can take the strain off of limited technical resources, allowing the developers on staff to focus on strategic needs. Additionally, processes that span the organizations (such as regulatory reporting and risk management) can be aided by Gen AI, ML, and natural language processing (NLP).

The Ongoing Evolution of AI for Payments

Just as customer needs will continue to evolve, so must banks’ approaches to innovation, product enhancements, and investments in payment infrastructure modernization. Harnessing AI technologies for payments requires establishing an appropriate data architecture, supported by effective data strategy and governance within the bank. Organizational culture, skills, and structure will also prove to be essential in order to optimize the use of rapidly-evolving AI, prioritizing use cases, while keeping up with customer needs and regulatory considerations.

Kieran Hines is a principal analyst in the banking practice at Celent, a global research and advisory firm focused on technology and business strategies in the financial services industry.