I recently listened to Dave Birch telling the audience the story of Frederic Tudor, the 19^th-century “Ice King” of Boston. The short version is that Frederic made his fortune shipping huge blocks of ice (we’re talking hundreds of tons) around the world in the pre-refrigeration era. He was constantly improving his methods of shipping and storing ice and was even an early adopter of refrigeration techniques to keep his cargo cold.

What Frederic didn’t foresee was just how fast consumers would adopt refrigerators and ice-makers once they were commercially available, and his business collapsed.

The parallel that Dave drew is the way that banks have been relatively quick at adopting AI for their own internal uses, and his cautionary tale serves as a warning to the banking industry, not to underestimate just how fast their customers will adopt AI.

I couldn’t agree with him more!

It is clear to me that the banking industry is on the cusp of a profound transformation, not only from banks using AI, but from their customers using AI (bots) to navigate their financial lives. I foresee that in the not-too-distant future, banks will be dealing more with bots than with humans …….. are you ready for this?

Kirsty Rutter (Group Strategic Investment Director CVC, Lloyds Banking Group) and Dave Birch (Principal, 15Mb Ltd) co-authored an excellent paper: ”Digital banking in the artificial intelligence era: Strategies for serving nonhuman customers”, which Kirsty shared in a LinkedIn post here if you are interested in getting your hands on it (after reading to the end of this post of course 😇)

What follows is my precis of their paper, along with a few additional points of my own.

The Rise of the Bots: A New Kind of Customer

Imagine a future where your customers aren't just individuals, but sophisticated AI agents acting on their behalf. These "customer bots" will possess an intelligence and speed far beyond human capability, making financial decisions, executing transactions, and performing analysis with unparalleled efficiency. This isn't science fiction; it's an inevitable progression driven by the interplay of open banking initiatives and increasingly "agentic" AI systems. These bots will be capable of independently carrying out complex, multi-step workflows without constant human oversight.

This means banks will increasingly move from a Business-to-Consumer (B2C) model to a Business-to-Robot-to-Consumer (B2R2C) model. The traditional customer journey, where humans interact directly with banks, will evolve into one where customer-controlled bots engage with bank Application Programming Interfaces (APIs).

The Threat to Traditional Bank Profit Pools

The transition to serving nonhuman customers poses a significant threat to retail bank profit pools. Historically, banks have benefited from customer inertia: the tendency for individuals not to switch accounts or refinance loans even when it's financially advantageous.

In short, we are lazy (in the most part), and switching is a tedious task.

In a world where bots are constantly seeking the best financial outcomes for their human principals, this inertia will disappear. Bots will relentlessly compare prices, fees, and rates, ensuring that their users always get the most favourable terms.

This necessitates a fundamental rethink of retail financial services engagement strategies.

The Shift to Financial Health Services and the Role of Regulation

One way to refocus strategy is by framing it in terms of financial health. Regulators are already moving in this direction. In the UK, the Financial Conduct Authority (FCA) has introduced a Consumer Duty of Care, requiring financial institutions to prioritise customer needs, help them achieve financial objectives, and ensure product suitability, value, and responsive service. This duty also includes preventing foreseeable harms, such as financial vulnerability.

The reality is that financial literacy is generally poor, and the financial landscape is complex. Even a basic bot, given access to bank accounts and operating under a duty of care, could significantly help customers manage their money more effectively by optimising cash flow, debt, savings, and investments. Just as a corporate treasurer manages an organisation's financial health, this agentic technology could offer individuals the same capability.

This aligns with a next-generation narrative centred on embedded finance, open data, smart wallets, and digital identity, all designed to deliver holistic financial health.

Strategic Elements for the AI Era

Kirsty and Dave suggest that to thrive in this new environment, banks must address three significant strategic elements: interfaces, wallets, and literacy.

1. Interfaces: APIs Not Brands

In the age of bots, the "face" of the bank will no longer be branches, websites, apps, or call centres, but rather its APIs. Customer bots will choose financial services based on technical factors like API functionality, open finance interface availability, service uptime, security, and data accuracy and richness, rather than emotional appeal or brand recognition. This means banks will need to compete much more sharply on price, leading to an inevitable "race to the bottom" if price is the only determinant.

To gain a competitive edge beyond price, banks must optimise their operations for bot preferences. This requires a fundamental rethink of existing processes, moving away from outdated batch processing to leveraging networked, cloud-based data and agentic technologies.

2. Wallets: The Crucial Enabler

Digital wallets are poised to become the central mechanism for managing transactions in the future. These will be "smart wallets" with integrated intelligent agents performing financial tasks that are either too tedious or too complex for most humans to handle. The Open Wallet Foundation (OWF) envisions these wallets as containers for digital assets, credentials, tickets, and keys, working in tandem with an agent software component that can add, remove, or process items.

For these wallets to be truly effective, they must be highly substitutable, allowing users to easily transfer credentials and assets between different wallet providers. Initiatives like the European Digital Identity (EUDI) Wallet Architecture and Reference Framework and the OWF are working towards secure and interoperable multi-purpose wallets, with a strong focus on digital finance and strong customer authentication.

3. Literacy: Bots as Financial Guides

The concept of financial literacy will be redefined. Rather than expecting individuals to become financial experts, the industry can leverage mass intelligent personalisation through bots. These bots can deliver financial education and insights in a way that is understandable, timely, and tailored to the individual.

Kirsty and Dave suggest that the scaling of basic financial literacy bots will be significantly easier than the widespread adoption of self-driving vehicles, suggesting that financial institutions should prioritise the nonhuman customer in their strategies now. This vision includes controlled data sharing between specialist bots to achieve financial health for customers, operating within frameworks like the EU's Financial Data Access (FiDA).

When price becomes less of a differentiator, product design can shift towards "belief-based" propositions. Customers may choose bots that align with their ethical perspectives, favouring sustainability, corporate social goals, or charitable status.

What Else is Needed?

Adding to the points raised in the paper, I believe that while APIs remain a necessary foundational element, current APIs have some limitations in a world of Agentic AI.

Most APIs today require manual discovery and integration, whereas agentic systems need machine-readable, self-describing environments that enable dynamic discovery and composition. Most APIs are also stateless, whereas agentic AI operates contextually and continuously, which calls for conversational protocols, memory and reactive feedback loops beyond the typical request-response paradigm of today.

This brings us to agent-ready protocols.

New protocols are emerging to allow agents to discover and compose capabilities dynamically. They support negotiation and cooperation between agents and services, and they define “intent-centric” communication instead of endpoint-driven interactions. These include the open source Open Agents Protocol, Microsoft’s open source AutoGen framework, LangChain’s LangGraph and many others currently under development.

Furthermore, for customer agents to be successful in the future, they will need to have the ability to “compare apples with apples” as it were. Yes, I’m talking about industry standards ………

A future-proof financial services system also requires standardised ontologies and intent frameworks.

Agents will need to leverage common / shared ontologies for financial objects and actions, complete with intent taxonomies that allow them to match their customers’ goals (aka YOUR customers’ goals) to capabilities across institutions. And of course, these need to be aligned with risk frameworks and regulatory constraints.

Good luck with that I hear you say!

But if your bank doesn’t participate in the ecosystem, will my bot simply ignore your bank and what it has to offer in the future?

Rethinking Relevance

Historically, banks have benefited from high barriers to entry and customer inertia, with a significant majority of European bank clients maintaining the same primary account for five years or more. However, this loyalty is already showing signs of weakening in some regions, such as the UK, due to easier switching schemes and incentives.

The bank of the future will no longer be able to rely on this inertia.

The future battleground for customers will be determined by which bot they choose to work with, not which bank. Consumers will select bots whose moral and ethical frameworks align with their own, and which bots will best help them achieve their goals. This will lead to a new ecosystem of customer bot agencies representing diverse customer profiles, needs, and behaviours.

For banks, the crucial questions will become: What is the role of the bank beyond product supply? And does the bot become the bank's new face in the market?.

The implication is that banks may need to provide bots that operate on behalf of customers, earning trust through regulation and a clear duty of care.

Conclusion

The AI era ushers in a new paradigm where nonhuman customers, powered by sophisticated AI agents, will reshape the financial landscape.

Retail banks must move beyond simply enhancing existing processes with AI; they must fundamentally rethink their strategies from the ground up.

By focusing on agent-ready protocols, API-first services, embracing smart wallets, and understanding the role of bots in delivering financial health, banks can navigate this revolution and secure their place in a future dominated by intelligent agents.

Scenario planning for this future is not just advisable; it's imperative.

The time to adapt is now!

Rik questions Andrew on his opinions in this "behind-the-scenes" Director's Cut debrief session, which shares a few more interesting insights and thought-provoking points for our audience. Watch on YouTube below, or listen on your favourite podcast platform here.