ROMA (ITALPRESS) – In three years, from the launch of ChatGPT to November 2022, the Generative Artificial Intelligence has attracted the attention of the whole world. Since then, the evolution of AI has continued without exhaustion, outlining an ever greater centrality of new and more evolved forms of artificial intelligence, in particular the nascent IA agents.
The financial services industry is today on the threshold of what many analysts have no doubt about defining as the “third wave” of banking information. After the digitization of core processes in the 1990s and the explosion of mobile and cloud channels in the last decade, 2026 marks the passage of artificial intelligence to new frontiers of autonomy and ability to act on its own. It is explained in a recent study by Fucino Digital – the section of the Banca del Fucino dedicated to technological and digital innovation – by Professor Gianluca Duretto of the University of International Studies of Rome.
The Agentic AI is designed not only to respond to an input, but to plan and act autonomously in order to achieve complex objectives. Compared to the old “rigid” automation of traditional software, these new IA agents manage workflows made up of multiple related objectives. However, man has the fundamental role of supervisor of the process in the most crucial points. But how to control this aspect?
Precisely the subject of supervision assumes a central weight in the use of these systems within banks. In Europe the regulatory scenario is guided by the EU AI Act: from August 2026 strict rules will become binding, which classify many banking applications as high-risk systems. It will be mandatory to record any automatic decision taken by IA agents, to ensure rigorous security protocols and to insert protective mechanisms that allow human supervision even in a highly automated process. The human hand must always be able to intervene if necessary.
A Boston Consulting Group and OpenAI report estimates that the economic impact of Agentic AI will be very significant: the systematic adoption of autonomous agents could increase the profitability of banks by 30% and reduce their operating costs in a range between 30% and 40% by 2030. Moreover, according to NeuronsLab’s research, the adoption of AI agents could improve the EBITDA of the average bank of 5.4% per year.
The main advantages for banks and their customers are concentrated in four areas. Prevention of fraud: IA agents adapt in real time to new criminal tactics and will soon be able to automatically manage false positives, reducing the time spent for each individual investigation by 50% and providing customers with enhanced security. Safety, controls and procedures Know Your Customer: The remote verification procedures of the identity for access to accounts and those related to money laundering risk controls see the completion times even fall by 90%, with a significant reduction in costs. Financial advice: We talk about systems that monitor customer financial portfolios 24/7 and reduce search times by 40-50%, making high-profile consulting accessible to a much wider range of customers. Credit access: IA agents independently analyze balance sheets, account extracts and business plan in any format, reducing customer response times in case of credit request.
“Naturally, the road is not without obstacles, including cultural resistance to change, the shortage of specialized talents and the presence of isolated data in old business systems – emphasizes the study of Fucino Digital -. Without forgetting the difficulties of adapting to a rapidly changing regulatory framework. In conclusion, the bankof the future will not simply be the one with the application for smartphones more appealing or with a prompt use of AI agentsc. It will be able to find the right balance between the autonomy of systems and human control, always putting the safety and the needs of the customer at the center of attention”.
– Press Office photos Banca del Fucino –
(ITALPRESS).
