A 'double-flywheel' effect is about to transform financial services

• Key insight: The application of AI agents operating on new payment rails and the rapid adoption of tokenized assets will work together to drive changes to the banking industry at an unprecedented pace, scale and breadth.
• Expert quote: "We overestimate the impact of technology in the short-term and underestimate the effect in the long run." —Roy Amara

This is the second article in a three-part series. The first piece explored how the combination of digital assets and AI agents are going to rapidly transform the banking industry.

Processing Content

In more than two decades of working in payments and transaction banking, I have not experienced change as profound and as rapid as what we are facing now. The pace, scale and velocity of innovation that the industry is experiencing are unmatched. I see two reasons for this.

The first one is the combination of AI and digital assets that is creating a perfect storm of conditions for ever-increasing levels of automation, efficiency, effectiveness, and speed. Agentic AI will serve as an accelerant as the economic actors of the future are not likely to be human, but rather bots. AI agents will require rails that are fit for purpose: rails that are programmable, always-on, and frictionless. These automated agents will increasingly and systematically drive financial transaction traffic toward the rails that are the most efficient and effective; and as a result, drive adoption without requiring humans to adapt their behavior. In innovations of the past, the main limiting factor in the adoption of new technology has been the relatively slow pace at which humans adapt their behaviors. The rapid cycles of tech innovation usually outpace human capacity to change. But I often wonder to myself, "What if humans are not the ones making the decision on which product best meets their financial needs?" AI will seek solutions that are better, faster, available, cost effective and seamless.

In addition to AI, there is another force driving more rapid adoption of tokenization, and that is a flywheel effect between payments and capital markets. It is not just about tokenization of money or tokenization of other assets as siloed and stand-alone innovations. The tokenization of both money and real-world assets, or RWAs, are mutually reinforcing and will put pressure on both sides to evolve and keep up. The promise and benefits of tokenized money and tokenized RWAs are not complete without each other, and this will create mutual pressure for faster development and adoption of tokenization across the full value chain in financial services.

While the two forces above are distinct, they will have compounding effects and drive greater levels of innovation and adoption. This double flywheel effect is why I believe the level and pace of change is unmatched.

In addition to the speed of change, I often think about the scale and breadth of the change. I often hear questions on the value of tokenization, and hear versions of the statement that digital assets "are a solution in search of a problem." Or that "clients are not asking for this."

My first reaction to those statements is that many financial processes are inefficient, error-prone, expensive and filled with friction; and one does not need to go in search of problems. There are plenty of client problems waiting to be solved, and many problems that we have been unable to solve with legacy technology. We do not need to do much searching to find opportunities to deliver better solutions. And while clients may not explicitly ask for tokenized solutions, they seek simple, efficient and integrated ways to manage money and assets. Product leaders should anticipate and address client pain points, rather than wait for the client to define and ask for a better solution. If a client asks for a specific solution, it is because a competitor has already educated them on the potential of the solution; and as a result, the competitor has already set the standards and expectations. I would much prefer being the one to define the vision of what is possible in the eyes of a client.

My second reaction to those statements is that we may have caused our own problems when we talk about the promise of technology. Are we so enamored with technology for the sake of itself that we forget the problems it is meant to solve — or to explain it and make it tangible? In some ways, our own hype is crowding out real problem solving and damaging our ability to move innovation forward. This level of hype can lead to a loss of credibility. As with any product innovation, we need to keep ourselves grounded and intellectually honest. And that means answering three key questions: What is the client problem that we are trying to solve? Why is this a problem worth solving? and How do we solve it (and why do we have the right to be the ones to solve it)?

If we answer those questions, then the story about digital assets becomes far more compelling.

Another frequent topic of discussion is whether digital assets are a product or a rail. And my answer is that they are both — and they are neither.

Read more:

• 'We could lose our edge': Bankers on competing for digital assets
• How Barclays plans to grow its U.S.-based digital bank
• M&T says it's monitoring impact of AI agents on deposits
• Why Chase is rethinking workflows as AI adoption accelerates

"Both" because tokenized forms of money and RWAs enable new features and capabilities for clients (improved product), and also because they enable faster, easier and cheaper movement of value and assets (better rails).

"Neither" because framing it in this way is far too narrow and limits us in our ability to imagine a future that is vastly different from what we have now.

The analogy I think of is mobile phones. One could have argued 25 years ago that cell phones are a better product (a smaller, more compact, more accessible phone) and a better "rail" (we could make voice calls without being tethered to a wall). But describing cell phones as simply a better product or rail for making voice calls underestimates how they have evolved and the impact they now have on our daily lives and activities.

The problem with a product or rail view of innovation is that we are limited by the uses we are familiar with at that moment in time; e.g., our thinking 25 years ago that mobile phones were a better product and rail for making voice calls. I imagine that for most of us today, the use of cell phones for voice calls represents less than 10% to 15% of the total usage. And yet, how many of us could have imagined 25 years ago all of the apps, functions, capabilities and enablers that fill the other 85% to 90% of our usage? How many activities do we perform on our phones that have nothing to do with phone calls? How many industries, new offerings, and new business and revenue models have proliferated as a result of mobile phone technology?

This is called Amara's law. Roy Amara was a computer scientist who said, "We overestimate the impact of technology in the short-term and underestimate the effect in the long run." We get tied up in the short-term functionality and benefits, and lack the ability to comprehend the compounding effect: Each innovation leads to further innovation that would not have been possible without the previous one. It is relatively easy to understand immediate and visible benefits of new technology, but nearly impossible to imagine what it will be in the future after multiple compounded innovation hops. A physics professor Albert Allen Bartlett put it best when he said: "The greatest shortcoming of the human race is man's inability to understand the exponential function."

My guess is that the exponential effects of tokenization of money and assets are likely to lead us down a path we are not yet capable of envisioning. A future where new businesses, business models and revenue pools we cannot yet anticipate will exist, perhaps overtaking many of the products and solutions that exist in financial services today — similar to how other uses and products, made possible by mobile phones, have surpassed usage related to voice calls.