Top 10 Themes in Banking in 2025

January is a good month for predictions and focus lists. Forbes produced a list of top 10 themes in banking in 2025. I’m not sure I agree with the Forbes article – so here is mine:

1. Operational automation steps up BUT the approach is still too much like RPA – the real value lies in reengineering with an automation mindset using composable techniques, much of which could be described as AI but the term is too broad. I think the whole ‘AI’ term will become tarnished, those that really work with it will talk about the application of more specific techniques.
2. If the US dilutes/rolls back some of the recent regulation, what challenges does that bring? A resurgent US banking sector increasingly dominating a reg-bound European approach. The cost of regulatory compliance and the actual value of the regs themselves will come under scrutiny in the US.
3. Continued modernisation of key banking technology but using a more micro, more agile approach.
4. The cost of existing applications estates is a problem that will increasingly choke off change unless reversed. Technology budgets may continue to remain under pressure or flat, forcing technology to find ways to do more with less, which they will. This is a spend negative vortex but value creative positive spiral.
5. The Cloud gets cloudy – there are real and definite use-cases but the initial all-in approach is not working and we may see a roll back in some instances.

6. I sense the press is getting frustrated with GenAI but as the Economist says, quite a lot is going on ‘in secret’. People are starting to learn the limitations of tools like GenAI, there is still a lot of naivety around the difference between deterministic and probabilistic tooling. More will happen ‘in secret’ – I am hugely bullish about the potential of the range of IA tools.
7. Modernisation of lending. One of the last remaining bastions of highly-manual processes and unstructured data. Modernisation is not straightforward, lending often exists in many divisions with many different processes built up to service many different markets/customer types. Refer to 3. For how to succeed, the way NOT to do it is a giant waterfall based consultant-fest.
8. Payments modernisation continues. Some of this is driven by changes to Swift which will lead to a general modernisation of bank payments infrastructure – much of this is inflight. Where this leaves the multitude of payments fintechs is highly questionable – are we about to see the disintermediation of these payment apps either by the banks or by ApplePay/GooglePay?
9. Data centres, the cost of processing and storage, the need for specific AI training and the availability of power causes a rethink of approach in the big banks. This ultimately could be good for the cloud providers as firms rethink the way they work with their infrastructure availability. This may seem inconsistent with 5 but it isn’t. The Cloud becomes more about infrastructure planning with respect to application and data usage patterns, the challenge will be to build highly dynamic storage approaches as code and embed/retro-fit these into application code.
10. Breaking Down Organisational Silos: The Rise of Data Mesh and Data Fabric: The continued deployment of data mesh and data fabric solutions is reshaping how firms tackle the challenges posed by Conway’s Law, where organisational silos are mirrored in technology and data structures. Previously, even when cultural barriers were addressed, technological limitations in scaling and integrating data systems hindered progress. However, advancements in data architecture, combined with the elasticity of the cloud, are breaking these boundaries. With technology increasingly able to handle large, decentralised data systems, the focus now shifts squarely to overcoming cultural silos and aligning organisational mindsets.