Tom Merry wouldn’t class himself as a hardcore technologist. As Accenture’s UK managing director for banking strategy, you won’t see Merry’s fingers bleeding from typing out endless lines of code or furiously pulling out and reconnecting ethernet cables in poorly-lit server rooms. But as Merry explains to Tech Monitor, he knows “enough to be dangerous”, a playful way of letting us know he’s privy to most of the hopes and fears of British banking CIOs about the latest trends in digital technology. In the following interview, edited and condensed for clarity, Merry walks us through how the UK banking sector’s C-suite is grappling with generative AI – and how some of the more established players in the market are planning on stealing a march on their more nimble fintech cousins.
When you’re talking to senior leadership figures in the UK banking sector, what issues around digital technology tend to be front of mind for them?
Tom Merry: I’m seeing this as an era of convergence in banking in the UK market. If you read through the strategies of all the banks, they all have the same goal, whether they’re a large incumbent high street bank, a digital-only neobank, or a more traditional challenger that sits somewhere in between. That target is to become an organisation that’s completely customer-centric and powered by next-generation technology which makes them more responsible, more relevant and more efficient.
That’s the utopia. But, of course, the journey to that point is different depending on what type of banking operation you’re running. Working out what that path looks like is the challenge for your average CIO. How can their organisation accelerate out of decades of slow transformation and go further and faster from an AI and data perspective?
Devising that strategy is more difficult for your larger, incumbent banks than the other types. That isn’t to dismiss how far they’ve travelled on their digital technology journey compared to 30 years ago. Back then, the cost of data storage and computer processing power was prohibitive. So, what did they do? They created large mainframe technology systems that captured a minimum amount of data, because it was expensive to store, and processed it as infrequently as they could.
That’s the bedrock upon which large incumbent banks have built their technology ecosystem. What’s ironic is that the situation that created those large mainframe systems has completely disappeared. Computer processing power is now so cheap that it’s perfectly possible to deliver real-time, dynamic insights to the customer. And yet, big, incumbent banks remain heavily hardwired into batch-oriented, big, large, monolithic bits of architecture. To really unpick that safely and securely and make progress away from that is something that remains a really big challenge.
They have made some progress. Established players are finding ways to creatively use next-generation technology in certain areas: they’ve done a huge amount, for example, to digitalise their front end. They’ve also made key investments in the provision of API-enabled architectures and committed wholeheartedly to cloud computing spending. In so doing, many UK banks have built the kind of ecosystem where a whole new wave of generative AI applications can really make a positive difference.
You’ve previously said that “AI will touch almost everything that goes on at a bank.” What kind of difference is generative AI making right now in UK banks?
Currently, CIOs are starting to see value in using these applications to support ‘business-as-usual’ functions: things like transcription, document summarisation, code generation, software testing, IT security, and workflow management. It doesn’t extend to more mathematically-based tasks, and the assumption that it can is something we’ve had to push back against. I’ve had at least one chief investment officer say to me, “Tom, it frightens me to death that you’re suggesting we might use gen AI to do future cash flow predictions for our clients.” I had to say to him that we definitely weren’t and that he might want to consider predictive and prescriptive AI for that task – which, as it turned out, his bank already was!
Despite that, I can still envision lots of opportunities to deploy generative AI across your typical bank. If deployed in a customer service setting, you can use it to personalise interactions; in marketing, you can use it to hone how your firm interacts with your customers on social media. You could also think about deploying it in a human resources context, to help optimise onboarding, for example. Client-facing employees like mortgage or wealth advisors, too, could use it to augment their responses to important questions in the moment and give them much more profound insight into a problem than they might have had before.
That leaves the C-suite asking many questions of themselves, starting with where they actually start? Are we capable of deploying these applications? Are these many use cases for generative AI actually desirable, and do they solve problems that exist for us? Fundamentally, does investment in generative AI align with our broader strategic goals and therefore represent a solid investment? Most of my clients are excited by what it might do in the endgame but, for now, are proving very careful and considered about where, when and how to deploy the technology in their own businesses.
Earlier this month we saw Barclays cut some 5,000 jobs in a new efficiency drive, primarily in the back-end of its business. How do you see greater adoption of AI within the industry influencing the scale and character of these restructures?
Whilst I won’t comment on specific firms, AI will have a great impact on ways of working in the banking industry. What’s crucial for banks is to ensure their AI strategy places the workforce and skills at its core.
The way we see it is, AI will augment or automate certain tasks, rather than eliminate entire job roles. As an example, our research showed that 39% of all work done at banks had the potential to be automated using generative AI, with 34% having the potential to be augmented with the technology. These are often tasks which still involve a crucial human touch, such as answering customer questions on loans or financial topics through an AI chatbot, analysing customer information to provide advisors with suggested next steps, or adapting the language or images of marketing messages to best suit the customer. This is where the potential in AI lies: helping people do more work, do the work better, and free them up from repetitive tasks.
What banks need to focus on is how they equip their people with the right training and skills to best leverage the value of AI and to do so in a responsible way. The banks that do this well will see an improved customer experience, a more efficient and productive workforce, and increased revenues.
We also saw HSBC launch its ‘Zing’ app to rival Revolut. Should we expect other, more established banks to launch similar products to steal a march on their fintech rivals?
To go back to my earlier convergence theory, every bank, whether that’s a large incumbent, a neobank, or a traditional challenger bank, is striving toward the same digital end game, with the aim of achieving the most profitability and healthiest balance sheet whilst doing so. So, it’s no surprise that incumbent banks will be exploring how they can compete with more specialised fintech propositions whilst they leverage the benefits of their large customer base that holds greater deposits.
Read more: ‘We had 45 different back-end systems’: Heathrow Airport’s digital transformation explained
