Week after week, with the quiet regularity of a scheduled maintenance update, another bank somewhere announces the closure of yet another branch. The rationale is delivered in language so polished it has become almost liturgical: customers have migrated online, mobile banking now handles the everyday, and artificial intelligence stands ready around the clock to field any remaining questions. Physical real estate, the argument runs, is an expensive anachronism. Yet somewhere in that tidy cost-benefit logic, an older and more fundamental truth keeps getting filed away — that the relationship between a person and their financial institution has always been, at its core, a matter of trust, and trust is built face to face.
This is the argument advanced by Chris Skinner, one of the financial industry's most prominent commentators, whose work at The Finanser has long tracked the tension between technological innovation and human-centred banking. Skinner's framing is disarmingly simple but points to something the sector's spreadsheet models consistently undervalue: eye-to-eye interaction — what he shorthand labels I2I — is not merely a delivery channel preference. It is the mechanism through which financial trust is established, tested, and renewed. Replacing it with a chatbot or a large language model, however sophisticated, is not a lateral substitution. It is a category error.
The Economics of Closure Look Compelling — Until They Don't
To be fair to the banks accelerating these decisions, the financial logic is not without merit on its surface. Branch networks are capital-intensive: leases, staffing, security, compliance, accessibility upgrades — these costs compound year on year. McKinsey and others have documented how the cost-to-serve a customer digitally can be a fraction of what it costs through a physical interaction. Mobile banking adoption has risen sharply across virtually every demographic, including age groups that were once reliably branch-dependent. And AI-powered customer service tools genuinely have improved — capable of resolving a large proportion of routine queries without human escalation, at any hour of the day.
The problem is that banks are making a mistake common to industries that conflate efficiency with value creation. Cutting a cost line and improving the customer relationship are not the same operation, and in financial services they can be directly opposed. A mortgage applicant navigating a difficult income situation, an elderly customer uncertain about a suspicious transaction, a small business owner structuring a first loan — these are not routine queries. They are moments of vulnerability, and vulnerability is exactly when the absence of a human presence is most acutely felt. No large language model, however well-trained, replicates the experience of a trained human being who can read hesitation, build confidence, or simply confirm with a look that a person's concerns are being taken seriously.
What the Data Misses About Human Behaviour
Banking institutions are sophisticated data collectors. They track login frequency, transaction volumes, app engagement, and support ticket resolution rates. What is considerably harder to quantify is the slow erosion of loyalty that occurs when a long-standing customer discovers their nearest branch is gone. Nor does any dashboard capture the new customer who never opened an account because there was no local presence to walk into when they were ready. These are the counterfactuals that never appear in the closure business case, but they are real economic costs — expressed in churn rates years later, or in the quiet migration of deposits toward institutions that maintained a physical footprint.
The competitive implications deserve scrutiny. Revolut, Wise, and the broader neobank cohort built their propositions entirely on digital convenience, and they have captured significant market share on that basis. But the established banks closing branches are not primarily competing with neobanks for digitally native customers — those customers are already gone. They are making decisions that affect the portions of the population for whom digital-only is not yet a comfortable default: rural communities, older individuals, people with accessibility needs, and anyone whose financial situation is complex enough to benefit from a conversation rather than a FAQ.
Trust Cannot Be Automated
There is a deeper philosophical point embedded in Skinner's argument that the industry would do well to sit with. Artificial intelligence is a tool for processing and responding to information at scale. Trust is something different — it is an emotional and relational state that accrues through repeated, authentic human interaction. Banks were, for most of their history, local institutions. The manager who knew your family, your business, and your circumstances was not a luxury feature. That manager was the product. The decades-long process of standardising, centralising, and digitising banking has stripped that model down to its bare transactional bones, and AI-driven branch closure is the final step in that direction.
The question worth asking now is not whether banks can afford to keep branches open. The question is whether they can afford the long-term reputational and relational cost of closing them — and whether any algorithm, however capable, can compensate for the loss of a handshake, a conversation, or the simple act of looking a customer in the eye.
Written by the editorial team — independent journalism powered by Codego Press.