Between machines and people, the artificial intelligence that surrounds us
“Artificial intelligence won’t replace banks, but it will change the way they think about risk and customer relationships. Institutions that manage to balance algorithmic speed with human trust will be the ones defining the next generation of finance.” This is one of the most significant passages from one of the first international press articles that attempted to understand how the arrival of artificial intelligence would reshape banks’ processes, activities, services, and relationships. It appeared in The Economist in the fall of 2023 with a highly evocative title: “Finance and the machine mind.”
Pay attention to the dates, they’re never random. In the United States, around that time two years ago, OpenAI burst onto the scene with ChatGPT. In Italy the arrival took longer, but by the end of 2023 it was being discussed (and used) more widely. Let’s start from here: we often imagine immersive and predictive technologies, or machine-learning-driven or bot-based processes, as something still far ahead. And yet these technologies, including agentic and generative AI, have already long been part of banking and financial organizations.
The work… increased
When Clara, a startup founded in Mexico in 2020, decided not to simply offer a corporate card or a digital account, but to rethink how Latin American companies manage expenses, it truly set out on a journey toward the future. Clara, featured in The Financial Times just a few months ago, built a platform combining physical and virtual cards, expense management, reimbursement automation, financial reporting, and internal workflows, all in a single environment easily accessible to SMEs in Mexico, Colombia, and Brazil. In 2025 the startup raised an additional $80 million to expand, obtained licenses enabling banking operations in Brazil, and became a unicorn with valuations reflecting not only its economic value but also its disruptive impact on financial services.
In the quiet-seeming digital desks of Latin America, artificial intelligence is no longer a promise repeated at conferences: it is a daily engine. And if there is a name that represents this shift, it is this fintech that turned the old concept of a corporate card into a data-driven spend-management platform. Today it manages physical and virtual cards, payments, and expense reports for more than 10–12 thousand companies across Mexico, Brazil, and Colombia, with over 7 million annual transactions worth more than $1.5 billion. It has joined Mexico’s unicorn club and has been included in Fintech Innovation 50 as one of the world’s most interesting corporate-payments companies. The context says a lot. Mexico has become one of the continent’s fintech epicenters: almost a thousand initiatives, including 773 local startups, annual growth close to 20%, and more than 70 million fintech users in 2024, a number expected to reach 86 million by 2027. The market for cards and payments already exceeds $215 billion and is growing at an annual compound rate of 11% until 2033, while digital payments are estimated at over $109 billion in 2024 with even more dynamic prospects.
Within this raging river, Clara positions itself as intelligent infrastructure for businesses: AI models that detect spending anomalies, suggest savings, automate compliance controls; digital assistants that “read” invoices, business trips, marketing campaigns, and propose actions before the CFO intervenes. The company itself describes its platform as able to “optimize your spending program with AI and never miss a saving opportunity,” marking a clear departure from old-style accounting software. It’s no surprise that global media have presented Clara as a symbol of the new Latin American fintech wave: Reuters, for example, highlighted that in Brazil alone the launch of new payment accounts aims at volumes of 6 billion reais within a year, while Bloomberg defines it a corporate spending-management firm racing to consolidate regional leadership.
Turning AI from a slogan into a concrete lever for business and market growth.
This story is not a distant tale, but a tangible example of how a plural idea, centered on people and service, can change the game.
Beyond niches
To understand AI’s impact, we must go beyond the perimeter of our work. Never before has artificial intelligence reshaped our lives at work, at home, with friends, and in our free time. It has become a popular topic, far beyond innovation-oriented communities. But what makes the difference is the cultural awareness of the epochal transformation occurring in workplaces, in a logic of alliance between people and machines—in this precise order. Within banking processes AI is already infrastructure. In credit-risk management it powers more timely estimates of probability of default and expected loss, integrates textual signals and nonlinear indicators, and enables early-warning systems on retail and corporate portfolios.
In payments it enhances fraud prevention within milliseconds, with models detecting anomalous patterns and reducing false positives; in anti-money-laundering, network analysis and automated OSINT assist with alert triage and beneficial-owner identification. The point is not science fiction, but operations: AI already makes faster, more targeted, more effective what once required manual sampling. Thus new technologies can make controls quicker, cheaper, and more effective, provided risks and enabling conditions are properly managed.
In market desks and treasuries, predictive models support intraday cash forecasting, collateral optimization, and best execution; in market surveillance, systems read texts and orders to flag anomalous behavior. The international trajectory is clear. The Financial Stability Board has updated its view: rapidly accelerating adoption, with benefits in efficiency, compliance, and personalization.
Prediction and personalization
On the operational side, the key word is implementation: well-governed data, reproducible pipelines, model observability. This is how prototypes become services.
Public discourse converges: on one side the benefits—automation of customer support, more effective anti-fraud, streamlined processes; on the other side the risks of opaque models and technological concentration that may create critical points. The keyword is caution, combined with the need to raise the bar, experiment, and write new pages of the future. In banks’ daily operations the scope of use is expanding. Predictive models anticipate credit deterioration and activate respectful pre-collection; semantic engines allow natural-language querying of policies, minutes, contracts, providing valid answers. Product personalization becomes feasible. The balance point lies between ambition and responsibility. “The pace and scale of AI require new capabilities,” writes the ECB. It is up to banks to declare confidence intervals, measure error, update models, and explain choices. This way innovation stops being a promise and becomes maintenance of trust, the true capital to preserve.
Not only technology
At Sella, artificial intelligence strengthens the Group’s pioneering role while also deepening understanding of the phenomenon. But it is not only technology. The human component is essential. And this is a matter of culture as much as technology. Paying attention to weak signals that become strong decisions. Today forecasts live inside models that learn, connect heterogeneous data, and provide answers in real time. The question is no longer “whether” to adopt AI, but “how” to make it work well: measurable, traceable, compliant. “For complex and structured organizations, adopting artificial intelligence is not just a technological matter, but above all an organizational one. It’s possible to acquire the best vertical technical skills, but if the organization is not clear on how it wants to evolve and what changes it intends to introduce compared to the status quo, the benefits in terms of process and work acceleration will be limited. This is why at Sella our work has two sides: bottom-up, receiving project requests directly from business areas and services; and top-down, directing transformation from above, especially in some business domains,” says Valerio Cicco, head of the AI area within the Innovation team.
On the customer side, AI paves the way for hyper-personalized digital experiences, a goal that was extremely difficult to achieve with traditional technologies.
“Thanks to the possibility of interacting in natural language with one’s app, the experience will become simpler, more intuitive, and closer to customer expectations. This will be crucial, because expectations are changing rapidly: two years have passed since the GPT boom, when we all began to familiarize ourselves with AI, and as a result the service level we expect from a service company will grow exponentially. Companies that anticipate this change will have a significant competitive advantage,” Cicco adds.
The customer at the center
“Artificial intelligence supports business-area growth with a strong focus on the customer because it truly helps companies grow—especially when the customer is placed at the center. For many years now, AI solutions, especially what we call traditional ones, have supported many business areas, generating concrete benefits. Just to give an idea, in the last eight years we have developed and managed more than 80 AI models that allow us, for example, to protect clients from fraud or to suggest personalized solutions.” This is how Federico Angaramo, head of the AI Competence Center, describes the team of data scientists at the Sella Group who create and refine AI-based models and solutions. Their main task is to develop solid and reliable systems, following rigorous methodologies and the development standards adopted by the Group. How is AI implemented in project plans? “The AI Competence Center supports both internal projects at Centrico and those of other Group companies. With our specialized skills, we support colleagues in AI-related projects: first we analyze available data, then identify the most suitable metric to evaluate whether a model is truly effective and useful; then we conduct a series of experiments that are validated and then made available to those who need them, whether colleagues or clients,” Angaramo explains.
Recently, the arrival of generative AI has given an additional boost: today, customer-assistance tools have evolved and may become the gateway to future banking services. “From my experience, the arrival of generative AI accelerated adoption and understanding like never before. However, I believe that in the coming years we will see many concrete advantages, both for customers and for the company, thanks to the natural integration of more traditional AI models directly into daily processes. One of the most promising areas involves agents, intelligent systems that can work together and collaborate by leveraging LLMs’ reasoning capabilities. The real difference will be made by how these technologies are adopted: companies that integrate AI with methodological rigor and careful attention into their processes will have a major competitive advantage,” Angaramo concludes.
The rise of agents
“Artificial intelligence provides valuable support for automating tasks, improving products and services, and making them more personalized for the customer from traditional banking products like credit or payments to suitability checks on transactions, all the way to the fundamental realm of cybersecurity.” This is what Stefano Priola says, head of technology strategy at Centrico with a special focus on AI adoption. Here, artificial intelligence is implemented starting from strategic and project plans, putting thinking before technology. “As Centrico is a company dedicated to IT services, primarily core banking for the Sella Group banks, and BPO (Business Process Outsourcing), AI is adopted to reduce time and improve quality. Concretely, AI is used in software development to suggest code lines, provide documentation, and find errors or optimizations in existing code. In BPO, it helps automate manual and repetitive tasks, such as extracting data from documents or reconciling data across archives,” Priola explains. And looking ahead, the potential for adoption across all areas will generate concrete benefits for customers and companies within the Sella Group.
“Our strategic plans include adopting AI agents that interact with customers, employees, and directly with IT systems. AI agents, studied in universities for over thirty years but now rapidly advancing thanks to generative AI, are software components able to autonomously interact within their environment to achieve a goal assigned by humans. They are trained by humans through documentation and examples, just like you would train a new hire,” Priola says. But how do we ensure they don’t make mistakes? Priola has his view: “Just like with new hires, a very thorough process of training, testing, and supervision is needed, with complete control over their operations so that any mistake is caught immediately and corrected—just as you would with a new colleague. This gradual and controlled approach is essential for the reliability and safety typical of our work.”
A hyper-personalized future and trust
“Artificial intelligence is an advanced technology capable of analyzing large volumes of data and generating insights quickly and adaptively. It is not an end in itself, but a tool: its value emerges when it supports the response to real customer needs. That’s why the starting point is always understanding needs and goals, so that AI can be integrated purposefully. For a banking group this is a substantial shift: processes and services will be rethought in a data-driven and personalized way, transforming customer experience and innovation capacity. All of this must occur in full respect of transparency and trust, essential elements for strengthening the customer relationship.” This is the fundamental view of Alessio Damonti. Ultimately, AI promotes business-area growth with extreme customer focus. More than that, it truly puts the customer at the center.
“Thanks to its ability to analyze behaviors, preferences, and needs, AI makes it possible to anticipate expectations and propose personalized solutions, increasing satisfaction and trust. This does not just mean selling more, but creating lasting relationships based on value and transparency. Moreover, AI simplifies customer experience by reducing time and complexity and by providing proactive services that address real—not presumed—needs. In this way, business growth becomes the natural consequence of a customer-oriented strategy,” Damonti says.
What will organizations work on in the near future as AI becomes embedded? According to Damonti, three directions: Hyper-personalized experiences: moving from standard products to tailored solutions that anticipate needs and create proactive interactions. Intelligent processes: evolved automation and predictive analysis to make workflows more agile, reduce complexity and timelines, and improve service quality. New relationship models: transparency and trust will become central, with AI supporting more empathetic, data-based advisory services, especially in banking. “In short, the future will not be merely technological but customer-driven, with AI as the engine for innovating without losing human value.”
The new AI professionals
Thinking and building among machines, bots, and people. Even here, a consultative model is key for a data scientist: “AI is an evolving world that has seen enormous acceleration in recent months and years. Our tools have changed as well, keeping up with the latest developments is essential. We need to design tech solutions that combine AI with traditional tools. We develop AI solutions to support business and risk-management decisions, improve processes, and enhance user experience. We oversee the entire lifecycle: from data analysis to production deployment. For us, consultancy means supporting the transformation of business needs into technological solutions that combine AI and traditional tools. We also provide consultancy through training and by designing applications.” This is how Alice Campigotto, a 30-year-old data scientist at the Sella Group’s AI Competence Center (with the company for four years), describes her work. The new world shaped by AI passes through this new generation of professionals who design advanced solutions. “Whether it’s tabular, numerical, textual, or image data, AI algorithms learn from data, knowing them and handling them is fundamental. Not by chance, my job happens to be a data scientist. In my case, I don’t just use AI, I develop it. Here we write the code to analyze data and train machine-learning models that learn from them. We also integrate third-party solutions through carefully designed architectures.” And the greatest satisfaction in her work? Campigotto has no doubts: “Teaching about AI: explaining that artificial intelligence exists, what ideas lie behind the main algorithms, dispelling the belief that it is something magical and infallible, and instead describing how it can support us.” Once again, even when discussing sophisticated and immersive technologies, the real game is played around what must always remain at the center: the person.