Most financial institutions are using AI to do existing things faster. The ones that will matter are using it to become something different. The distinction is not technological. It is strategic.
Retail banking is no longer a strategic frontier. It is infrastructure.
Across global banking, billions are being deployed into artificial intelligence programmes that automate onboarding, reduce call centre volumes, and accelerate credit decisions. These investments are not wrong. But they are focused on the wrong layer.
Most institutions are using AI to perform existing tasks with greater speed. This delivers incremental efficiency, but it does not address a more fundamental shift in how value is created in financial services. The distinction between banks that survive and those that thrive will not be technological. It will be determined by where AI is applied, and what it enables the institution to become.
Horizontal efficiency is the application of AI to make existing processes faster, cheaper, and more scalable. It is the dominant focus of retail banking transformation.
This layer automates routines, streamlines onboarding, and reduces cost to serve. It is necessary to remain competitive. It is not sufficient to differentiate.
Retail banking now operates as a utility layer. Competing on experience alone is equivalent to competing on uptime in cloud computing. It is expected, but it does not create advantage.
The economics reflect this shift. Margins are compressed by regulatory burden, fee pressure, and rising expectations for seamless digital access. In this environment, the retail bank becomes a self-driving financial system where operations must be automated, low cost, and invisible.
At the same time, banks are making increasingly bold declarations of purpose. They speak of financial inclusion, sustainability, proactive fraud protection, and customer advocacy.
The structural reality does not support these ambitions.
Operationalising a purpose-driven organisation requires real-time data integration, advanced analytics, and decisioning capability that extends beyond traditional product silos. It requires sustained investment in both technology and talent.
Banks are attempting to fund this through a retail model that is structurally constrained.
The industry cannot close this gap through marginal efficiency gains. Purpose will not survive as a loss leader. It must be funded by a different engine.
Vertical innovation is the application of AI to redefine what decisions the institution can make, and therefore what value it can create.
This is where advantage shifts.
The margin that funds purpose lives in the complexity of business and corporate banking. This is the domain where problems are not standardised and cannot be fully automated. It is where judgement, context, and structuring capability determine outcomes.
In this layer, the role of the bank shifts from product provider to strategic partner.
A bank that can model second-order supply chain disruption in real time does not simply process trade flows. It can anticipate inventory shocks, adjust credit exposure before stress materialises, and reprice risk ahead of competitors.
A bank that can simulate capital structure outcomes across jurisdictions does not simply originate loans. It can design financing strategies that optimise regulatory treatment, liquidity position, and long-term cost of capital for the client.
The corporate franchise provides the fee-based income, balance sheet utilisation, and strategic relevance required to fund the institution's broader mandate.
To remain relevant, banks must execute a dual transformation.
They must compress retail into a low-cost, automated utility that delivers access, inclusion, and data at scale.
At the same time, they must concentrate capital and talent in domains where complexity creates advantage. This is where human judgement, augmented by AI, produces differentiated outcomes that clients will pay for.
This is not a workforce optimisation exercise. It is a redefinition of what the institution is designed to do.
Purpose does not sit alongside this shift. It is downstream of it. It is funded, or it is not real.
The future bank is not one that applies AI everywhere. It is one that applies AI where it changes what the institution is.
Efficiency is necessary. It is not sufficient.
Banks that continue to optimise the wrong layer will build highly efficient systems that no longer matter.
This is the kind of problem I work on directly. If your organisation is navigating AI strategy, operating model design, or the economics of transformation, it is worth a direct conversation.