I recently joined Neil Joseph on the Xpanding Horizons podcast to explore how artificial intelligence (AI) in finance is changing the structure of banking, from the role of financial institutions to the nature of money itself.
When I wrote The Great Transition, I argued that finance would move towards a decentralised model where individuals interact directly with each other. As I develop these ideas further in my forthcoming book, Building the AI Bank, the role of AI has become central to how that transition unfolds.
What we are seeing today is different.
Intermediaries have not disappeared. Even in decentralised finance, new layers of intermediation have emerged in forms that resemble traditional financial markets. The shift now is not being driven by decentralisation alone, but by AI.
AI introduces a different mechanism. It allows individuals to create AI agents that act on their behalf, interacting with financial systems based on user-defined instructions rather than institutional interfaces.
This changes the starting point of finance.
Today, the bank defines the relationship. In an AI-driven model, the individual defines it through agents. The role of banks shifts from owning the interface to supporting the transaction.
That shift removes one of banking’s core advantages: time. Transactions are expected to be instant, reducing the value of float and forcing banks to compete on speed, identity validation and real-time liquidity.
At the same time, parallel systems are gaining ground. Stablecoins and digital assets have found traction because they are simple, liquid and globally understood. They allow small businesses to transact internationally without relying on traditional banking infrastructure.
This places pressure on both banks and central banks. As alternative financial systems become more usable, control over money becomes less absolute and more dependent on trust.
The implications extend to financial products themselves. Credit decisions are moving away from historical data towards real-time behavioural inputs, much of which sits outside the banking system. Financial products become faster, more fluid and increasingly tradable.
What emerges is not simply a more efficient market, but a different kind of financial system.
We are moving from a market economy, defined by institutional control and information asymmetry, towards a networked economy, where information is more evenly distributed and relationships play a larger role.
This does not guarantee better outcomes. Access may improve, but pricing and incentives remain shaped by those who build and fund the system. Technology does not resolve questions of fairness in finance.
In that context, the idea of fintech as a separate category has largely run its course. Much of it has been absorbed into traditional banking. AI represents a deeper shift, one that challenges the definition of financial products, the role of institutions, and the architecture of finance itself.
The question is no longer how to improve banking.
It is how to operate in a system where transactions are autonomous, relationships are mediated by AI agents, and value moves across networks that no single institution controls.
That transition is already underway.
