Could AI Make Me Value My Bank More, Not Less?
Guest Editorial by San Nakra-Shah, Co-Founder and Managing Partner at ChilliMint Europe Limited
If I were running a bank right now, I think I’d be a little worried about AI. Not because of the technology itself, but because of what it could do to my customer relationships.
Imagine an AI that can constantly compare savings rates, mortgages, cards and other financial products, work out whether there’s a better option and eventually act on my behalf. It could make it incredibly easy for customers like me to shop around.
And I do shop around. My financial life is already spread across several providers because I choose them for different reasons. I have one provider I tend to use when I travel, another for points and rewards, while my mortgage, savings and current account sit elsewhere. I’m not sitting around waiting for AI to tell me I have choices, but it could make exercising those choices considerably easier.
I’ve written before that I don’t really think in terms of having a “main bank” anymore. I choose different providers for different reasons. But that doesn’t mean the relationship I have with those providers has no value. In fact, I wonder whether AI could make some of those relationships more valuable. You can see why AI might initially make banks nervous. If products become easier to compare and switching becomes easier to execute, surely the relationship between a bank and its customer becomes less valuable?
I’m not sure it has to.
What if banks used AI to make that relationship more valuable, not less?
My bank already knows a lot about me. It knows when my mortgage deal ends and when my fixed-term savings mature. It knows how much money I tend to leave in my current account, which products I hold with it and which benefits come with my packaged account.
Banks have spent years becoming increasingly sophisticated at using this kind of information. They segment customers, personalize communications and work out what we might need or buy next. There’s nothing wrong with that. Banks are businesses and, of course, they need to sell things.
But if you know enough about me to work out what you might sell me next, do you know enough to work out when you could help me?
Take my savings. If a fixed-term product is approaching maturity, the bank already knows. Of course it will communicate with me, but it could also help me understand my options, what happens if I do nothing and whether there’s something better I could be doing with that money.
If my mortgage deal ends in six months, help me start thinking about my options before I’m already shopping around. If I’m paying for a packaged account, make sure I’m actually getting value from the benefits that come with it.
None of this is particularly futuristic. Banks can already do some of it today, which is partly why I find the AI opportunity so interesting. It could allow a bank to do this much more intelligently and personally, across millions of individual customers.
Instead of only trying to work out the next best product to put in front of me, what if AI was also working out the next useful thing I should know?
That’s the bit about AI in banking that interests me. I don’t particularly want another banking chatbot. I don’t wake up wishing my bank had more AI, and I don’t really care what technology is sitting behind my banking app. I care whether my bank does something useful for me.
I’d be quite happy for my bank to notice something I might have missed, join a few dots and say, “San, here’s something I think you should know.” I can then decide what happens next.
That might sound relatively modest compared with some of the predictions around agentic AI, but customers allowing AI to do more on their behalf may not be that far away. The FCA’s recent Mills Review found that one in five UK consumers, around 11 million adults, say they are likely to use AI capable of acting autonomously within pre-set goals. The same work also highlighted concerns around trust and control, which I think are every bit as important as the appetite to use it.
For banks, I think that creates an interesting tension. A customer-side AI could make individual financial products easier to compare and replace. At the same time, a bank’s own AI could make the relationship around those products more useful and more valuable.
And perhaps that starts to change what we mean by a valuable banking relationship.
What makes a banking relationship valuable anyway?
Banks have traditionally looked at things like the number of products a customer holds, where their salary is paid or how much of their financial life sits with one provider. But I could hold four products with one bank largely through inertia and feel very little connection to it. Another provider might only do one thing for me, but do it brilliantly, regularly spot something useful and earn my attention when it matters. Which one really has the stronger relationship?
That matters commercially too. Banks still need to sell products. But perhaps the most valuable next action isn’t always the one that produces an immediate sale. Sometimes it might simply be doing something useful enough that, when I next have a financial decision to make, you’re one of the providers I want to listen to.
Over time, the bank could also learn what useful means to me. I may not want to be interrupted to save £3, but I’d want to know about £300. I might value convenience over squeezing every last fraction of a percentage point from my savings, while someone else might make exactly the opposite choice.
That feels much more valuable to me than personalization that simply gets better at predicting what I might buy.
If you know more about me, what do I get back?
And there’s another reason this is worth thinking about now. The FCA’s Open Finance roadmap envisages consumers being able to share their financial data more widely and securely, potentially giving providers a broader view of areas such as savings, mortgages, investments, and pensions. The first priority consumer use case is mortgages.
So, with my permission, the amount a financial provider could understand about me is likely to grow. And if I’m going to allow you to know more about me, I think it’s reasonable to ask: what do I get back?
I’d like the answer to be more than better targeted marketing.
Perhaps that starts with my bank getting really good at saying here’s something I think you should know. If it consistently gets that right and earns my trust, maybe eventually I’m comfortable with, would you like me to sort that for you?
I think that progression matters. Before I let an AI act for me, I’d want it to prove that it understands what useful looks like for me. That feels like a much more important step than simply launching an AI agent and expecting customers to use it.
There is plenty of discussion about AI agents eventually sitting between banks and their customers, choosing products and providers on our behalf. I can absolutely see that changing the dynamics of banking. But the more I think about it, the less convinced I am that AI automatically makes the banking relationship less important. It may simply change what makes that relationship worth having.
If AI makes individual products increasingly easy to compare, the experience and value surrounding those products becomes another reason to choose one provider over another.
I still don’t think I need a “main bank.” AI may make it even easier for me to spread my financial life across different providers. But perhaps that makes the relationships I choose to value more important, not less.
The question for banks then becomes less about how much of my financial life they can own, and more about how useful a part of it they can become.
If I were a bank, I’d spend less time worrying about AI making it easier for customers to leave, and more time thinking about how I could use it to give them reasons to keep choosing me.
About the author
Santosh "San" Nakra-Shah is the Co-Founder and Managing Partner of ChilliMint, a specialist consultancy and marketing agency focused on payments, fintech, retail banking and financial services.
With more than 25 years of experience across banking, payments and financial services, San is a trusted advisor to payment schemes, banks, fintechs and merchants, helping organizations strengthen their marketing strategy and positioning, engage customers more effectively and bring complex products and propositions to market through clear, compelling marketing and communications strategies.
Before co-founding ChilliMint, San spent 10 years at Barclays in a range of senior roles spanning retail banking, payments, innovation, product development, and customer experience. Her final role was Head of Payments Innovation, where she led the development of new payment propositions and emerging payment technologies. During her time at Barclays, she played a key role in launching several industry-leading initiatives, including text message banking, contactless payments, and prepaid products.
She also led the bank's internal design agency, giving her a unique blend of expertise across product innovation, customer engagement, design and commercial delivery.
