The NatWest CIO goes over how they’re using AI with some use cases. Some:
- “Today at NatWest, over 40% of all code is either AI-generated or AI-assisted.”
- AI use: “If you’re a relationship manager, if you’re a branch manager, or you’re sitting as a teller in the branch, whatever that role is, you’ve got better customer insight, you’ve got better information at your fingertips. You don’t have to fill in loads of forms and go through lots of governance processes because actually AI can do that for you. You can have a much more engaging relationship with your customers, and that is really important.”
- Taking care of internal paperwork: “It’s about how do we free up our people so that they can have more of their time to actually serve our customers, rather than having to deal with their own internal manual process.”
- Making a smaller, custom fit model to run locally(?): “The goal there is that we can customize it to be very specific, to solve specific problems, and we can run much smaller models – 7 million parameters, not 1 trillion parameters, models that can actually run on your device directly. You’ve got things running locally on your phone that can be super-quick, super-efficient.”
- Public cloud enterprise AI is expensive: “The frontier models that everyone uses, ChatGPT, Gemini, Claude, these models are incredibly, incredibly powerful, but they’re incredibly large. Huge, in fact – a trillion parameters. That becomes expensive to run, energy inefficient, huge trading costs, huge inference costs.”
Leave a Reply