リテールバンカーは行動メタバースを必要としているか？

The other day I came across a video of a Siemens CEO Dr Roland Busch delivering a keynote speech at CES, the major trade show organised by the Consumer Technology Association. In the video, Dr Busch described industrial metaverse, “a virtual world that is nearly indistinguishable from reality, enabling people — along with AI — to collaborate in real time to address real-world challenges.” In the place where physical and digital worlds collide, customers would “accelerate innovation, enhance sustainability and adopt new technologies faster and at scale, leading to a profound transformation of entire industries and our everyday lives.”

According to Siemens, there are three main building blocks at the heart of the industrial metaverse: digital twins, software-defined automation, and data and AI. Before investing into large-scale physical assets – airplanes, factories, and so on – companies can start by building digital twins of those assets, which can be designed, tested, and refined at a fraction of time and cost that it would take in the real world. Software-defined automation helps bringing the finalised digital twin into physical world. Most of the world’s automated factories today rely on small physical devices, programmable logical controllers (PLCs), that use the software and hardware inside them to act like the brains for various machines inside the factory. The next generation controllers are going virtual, which can be monitored, controlled, and reprogrammed remotely and without requiring knowledge of highly specialist programming languages. Finally, good brains rely on data. Apparently, a highly automated factory generates 2,000 terabytes of data each month, which is equivalent to 500,000 movies! Edge devices can capture that data, analyse, filter, and only submit what matters. And while Siemens already had a partnership with Microsoft on industrial copilot, at CES they announced a partnership with AWS that would make “generative AI more accessible to application developers through combination of AWS Bedrock and Mendix Low-Code platform.” Specialised GenAI models become accessible and can be included by developers in new applications with a few clicks.

All this got me thinking if a similar industrial metaverse would work in financial services. Of course, concepts like digital twins are not new, and various segments of FS are already exploring them, from insurance to wealth and asset managers. My colleague Karun Arathil wrote an excellent report on How Digital Twins Will Help Insurers to Optimise Operations back in 2022. However, Insurance is arguably closer to physical world than Banking: it cares about physical assets – planes, trains, and automobiles – and can use digital twins to better model and understand various risks against which their clients want to insure. The laws governing physical assets are in the domain of physics and are well understood.

Of course, Banking also has and cares about physical assets, from retail branches and headquarters to various plants and factories their clients want to build and require financing. However, fundamentally, banks' products and services are intangible, not physical. Retail bankers are starting to experiment with digital twins of their customers to try and model how a new product might be adopted or the likelihood of delinquencies under certain economic conditions. However, here the laws are different: we are in the realm of behavioural science, and while the science has been making huge progress, individual behaviours remain more difficult to model and predict.

So, does it mean that retail bankers should be thinking about behavioural metaverse, rather than the industrial one? I don’t know, it probably matters less what we call it, but the retail bankers certainly should be – and many are – thinking about how to turn the wealth of data they have about their customers into true insights and intelligence. They need to get better at managing all that data, and they need AI at their fingertips. And for that, they will need partners, from low-code platforms to payments vendors, and from cloud platforms to AI specialists.