Hype, hope or holy grail? Unlocking new opportunities in payments processing with AI

A recent independent survey conducted by Celent found that 73% of corporate banks have reported clear revenue benefits from investment in advanced data analytics. Looking ahead, analytics, intelligent automation, and artificial intelligence (AI) technologies are now leading the investment agenda.

To explore the potential to bolster banks’ revenues and margins, I recently joined Celent’s Principal Analyst, Banking and Payments, Kieran Hines, and MongoDB’s Field CTO Boris Bialek during a recent webinar to discuss key trends and considerations:

1. Generative AI – meet the need, not the hype

Given the wild success of ChatGPT, there can be little surprise that generative AI is firmly on the radar for banks. Around 55% are currently evaluating or testing generative AI in some capacity, while 36% think the tech will have the greatest impact on the market in five years.

Yet the first question any bank should be asking before progressing projects is “why do we need to do this?” If the answer is “because we can”, it is not smart strategy and banks should take time to fully understand the implications. For example, a relatively simple use-case of converting unstructured data into structured data (key for ISO 20022 migrations) requires massive computing power, which comes at a significant cost. This means banks should carefully consider their position in the value chain and how much they are willing to pay for each prompt – and direct their resources towards the use-cases that can deliver meaningful returns.

There are also the significant challenges associated with data quality. When output is only as good as the input, robust data principles and preparation are foundational. It also means, somewhat paradoxically, that after spending decades as an industry investing heavily in automation to mitigate and eliminate human error, we must also understand where to insert human guardrails back into the process.

Data privacy is another critical consideration. This is driving rapid development and experimentation across private large language models to help banks move faster, while managing sensitive customer data and ensuring compliance with stringent regulatory requirements.

2. Delivering for corporate clients starts with data fundamentals

Corporate clients are continuing to turn to their bank partners to help boost operational efficiency and more effectively manage the treasury function within an increasingly complex ecosystem. This is creating new revenue opportunities for the delivery of services that offer real-time data and value-adding insights. For example, 77% of corporate clients report they would pay for insights-driven tools and 64% would pay for real-time cash forecasts.

Conversely, corporates increasingly see offerings such as automation across payables, streamlined onboarding and virtual accounts as table stakes – creating the risk of churn.

While AI undoubtedly has the potential to augment the delivery of new and enhanced services, one-off initiatives and tactical product updates will not deliver meaningful revenue boosts or cost reductions. It may not be glamorous, but banks should initially focus their attention on executing the basics and delivering standardised, accurate and complete payments data in real-time. This then provides a foundation for innovation that can move the needle longer-term.

3. Solving the developer resource crunch

Software developer capacity limitations remain an ongoing headache. Among Tier 1 banks in Europe and North America, 45% ranked a lack of developer capacity as one of the three most important barriers they face, ahead of current technology limitations, compliance, and broader budgetary issues. Developer constraints meant that, on average, banks reported missing around four opportunities to launch revenue-generating enhancements to their payment processing offering over the past two years. This equates to an opportunity cost estimated to be around 5% of annual payment revenues.

As a result, banks are starting to think differently about the payments value chain and re-assess their approaches towards software development. Increasingly, this includes exploring the use of low-code tools and platforms for payments processing. It also raises the prospect of leveraging AI to enhance developer efficiency.

Yet we must consider that large banks increasingly want to take control of their payments processing platforms. They don’t want to buy a black box – they want the ability to manage, upgrade and maintain their own technology. This demands an awareness and understanding of the underlying code, limiting the appetite to offload this crucial component to AI.

A far more compelling opportunity is using AI to support rapid, robust testing of code to not only detect bugs and errors, but also identify support for key business cases.

Developing a meaningful AI strategy

You don’t need to have been in the banking industry for long to have seen the latest ‘next big thing’ come and go. While AI and advanced technologies do promise to support product enhancement and innovation, we should not overstate their potential. Revenue and margin gains will likely be realised through the delivery of incremental improvements, and will be predicated on wider investment to upgrade outdated legacy infrastructure.

It’s also important to recognise that the potential future capabilities of AI do not upend good business practices in the here and now. Banks must carefully evaluate the areas where AI can deliver value and understand the associated cost factors. As with any technological innovation, this requires a clear strategy, an understanding of the underlying architectural requirements, and alignment with wider organisational initiatives.

Celent’s report, ‘Harnessing the Benefits of AI in Payments’, is available to download here.