By Gisela Colombo
Origins and trust
In 1397, the Medici opened a bank in Florence that sold more than credit. It sold something harder to manufacture: personal trust. There were no public balance sheets or rating agencies. There was a man’s word, backed by a network of correspondents who knew one another, who had seen each other’s faces, who knew whom they were trusting with their money. The modern bank was born, to a large extent, as a technology for extending trust between strangers. And yet, six hundred years later, that original problem is still not fully solved.
The scene repeats itself today in any country in the region, albeit with different scenery. Someone notices a strange movement in their account, panics, opens the bank’s app, and the first thing they find is a chatbot offering five options that don’t describe their problem. There are no Medici, no correspondent, no one on the other end who recognises their voice. There is a decision tree that never anticipated the panic of a real customer.
Hard data
That contrast is not anecdotal. A series of surveys carried out this year in Latin America, Spain and the United States agree on something that contradicts the bet banking made over the last decade: automating every touchpoint. Technology solved the operational side. It did not solve trust, which remains, as in Florence, a matter between people.
The regional data are compelling. According to the Financial Health and Payment Behaviour Radar, a survey by D’Alessio IROL, 46 per cent of customers of banks and digital wallets in Latin America want a better human relationship with their financial institution. When asked what they would improve first, 39 per cent chose, above any other option, being able to talk to a person when a problem arises. “Customers distinguish between the problem they have and the experience they expect,” sums up Nora D’Alessio, director of the consultancy. And she adds the line that should be engraved in every banking innovation committee: technology solved much of day-to-day operations, but technology alone is not enough.
Spain tells the same story with different numbers. The study Perception and habits of Spaniards regarding the banking sector 2026, by Nickel, found that 76.5 per cent of customers believe their ideal bank should offer personal service when they need it, and 75 per cent demand it outright. Some 50.6 per cent reject being forced to book an appointment or to bank exclusively online. 41.4 per cent changed banks because of poor service, and 31.1 per cent left specifically for lack of in-person service. In the United States, 79 per cent of consumers prefer a human agent over an artificial intelligence assistant to solve a problem, according to Lorikeet data. But these numbers should not be read as a rejection of the machine: 51 per cent go straight to a bot when what they need is speed for a simple query. This is not a war between the human and the automated. It is about knowing, at every moment, which of the two should be on the other end.
That is where the paradox running through this century’s banking appears. The more sophisticated the technology a bank deploys, the more it depends, for that technology to work, on something no artificial intelligence can produce on its own: the sense that someone stands behind what is promised. The Medici knew this without needing an algorithm. A bank doesn’t lend money; it lends a promise, and promises have always needed a voice to back them.
This does not mean automation was a mistake. 92 per cent of companies that implemented artificial intelligence in customer service report improvements in user satisfaction, according to the same source that documents the 79 per cent preference for humans. Both figures are true at the same time, and therein lies the key almost no one wants to see: well-placed artificial intelligence frees up human time for the exact moment when that human time matters. Badly placed artificial intelligence eliminates it right there.
Some solutions are already built on that distinction, rather than ignoring it. Instead of replacing the account officer or the service agent, there is technology designed to give them back time and context: letting the machine cross-reference data, anticipate the problem and prepare the answer, so that the person can focus on what no machine can offer, which is listening and deciding with judgement. It is the same logic summed up by Julián Colombo, co-founder and CEO of N5, when he describes where the relationship between banks and customers is heading: “AI will optimise, but human connection will remain key.” He had previously put forward the other side of the same idea: “AI will democratise empathy in financial services”, not because the machine will feel empathy, but because it can give everyone at the bank the time and information they currently lack to exercise it.
Five hundred years after a Florentine banker learned that credit rests on people and not on balance sheets, the question every bank and every fintech in the region should be asking is not how much artificial intelligence to add. It is older than that: when, exactly, does their customer need to find someone on the other end who stands behind what is promised.
Sources: D’Alessio IROL, Financial Health and Payment Behaviour Radar, via Infobae. Nickel, Perception and habits of Spaniards regarding the banking sector 2026, via DiarioAbierto. Lorikeet, AI Customer Service Statistics. Julián Colombo, CEO of N5, via blog.n5now.com.

