Three Tech Takeaways from Sibos 2023

Sibos 2023 is a great gathering of the financial industry and the place to take the pulse of what is exciting and challenging people.  

While the conference stages capture the limelight with industry visionaries, it is the grassroot opinions of the delegates roaming the conference floor, the foyers and – yes – the bars that are even more insightful. 

It was telling that while AI was a major topic for Sibos delegates, there were other more here- and-now issues that came across as loud and clear. 

Not There Yet on ISO20022 

While the big banks are all mostly set, Sibos revealed to me how there are so many smaller, mid-sized banks and business clients who are not there yet on ISO20022 transitions. There is a good deal more effort that needs to be made in transitioning these organisations’ core payments platforms and exceptions processes. Larger institutions who are ISO20022-ready were hoping that Swift would stay firm on the 2025 deadline. But, a question that did come up was could there be a need for a postponement to help these other organisations catch up? That is hard to predict but most banks we talked to would rather the deadline stayed as otherwise budgets and work slips, leaving most thinking that 2024 has got to be the year this transition hurdle is jumped. 

Rise in Real Time Payments Breaking the Bank? 

From the delegates there was a clear message that volumes for real-time payments are really rising very fast. A report released by Cap Gemini at Sibos provided some significant numbers on this. They reported that global non-cash commercial payments traffic is growing at a compound annual rate of around 11%. The mood music at Sibos 2023 was that commercial digital payments were catching up and getting close to overtaking the consumer digital payments bandwagon.  

But the message I also heard at Sibos was that volumes of cross-border payments were leading to pressures to again improve service levels in this area, but more broadly too. Banks’ commercial clients want the same payments experience they enjoy for their own personal finances. This means greater transparency and flexibility in how payments are managed, despite how commercial payments retain many manual processes and interventions. The room to keep pace with these pressures is more constrained now by rising inflation levels, reduced business, and pressures on budgets, making “big bang” strategies to undertake wholesale transformation of payment infrastructure harder to achieve. Could this create the ideal conditions for more agile approaches to digitise and automate payment flows and exceptions?  

And of course, Gen AI was a Big Thing at Sibos but … 

As I, and probably every other person predicted, we talked about Gen AI a lot at Sibos. It was great to get feedback on what the real use cases could be for commercial banking. There were two big fields of interest.  

One was how generative AI could help non-software coders speed up design and development of new apps and processes that commercial banking teams want. This is very much how generative AI makes the low code application development model even more useful. An injection of gen AI into these tools lets a team whip up a workable prototype much faster and move more quickly to being able to streamline processes around areas like payment repairs or trade finance or lending.  

The other is how Gen AI can help teams work faster and more efficiently within quite complex and pressurised workflows. Sibos had many great examples of how this flavour of AI could be useful. For example, we shared a use case that translated dense blocks of Swift messaging into succinct plain English that was instantly understandable and could be quickly shared on an email or used by the front office when talking with clients.  

Yet, these kinds of applications also demonstrated how gen AI is great but limited if not combined with process AI. Generative AI can summarise a situation but the technology itself is not adaptive and generates an answer simply from the data it accesses. By contrast, process AI can be self-learning and has predictive models built within it. Based on this, there was a dawning realization of the importance of how process AI steps in and helps humans apply AI to sort out issues or at the very least guide them on the best path. 

So, Gen AI caught people’s attention at Sibos but the reality of what commercial banks and clients do post- Sibos in terms of technology will be more about a continued focus on operational and service improvements, being able to deal with volumes through automation in areas like sanctions, improving self-service options, creating an omni-channel service experience, and coping with real time. What is exciting is how all of these challenges can be solved by elements of AI alongside other automation technologies and imaginative approaches.