At the International Conference on Sustainable Computing and Communication Technologies (ICSCCT 2026), held by the Faculty of Information & Communication Technology at the University of Malta and published in collaboration with Springer, Subhankar Panda took the stage as a speaker to make a case that few people disagree with but many are slow to act on: based on the next decade, banking education will be chosen by the next decade. their back-office tools.Subhankar’s talk went beyond the popular narrative of AI as a cost-cutting feature that is embedded in embedded systems. Instead, they created AI and ML as tools that should be part of the bank’s financial planning function itself – shaping how risk is affected, how money is structured, and how decisions that used to take weeks of committees are now based on model minutes.“Banks that take AI as an extension will continue to optimize yesterday’s experience. Banks that take them as partners are starting to ask good questions about tomorrow’s money.”From automation to anticipationA recurring thread in the address was the move from AI that performs known tasks to AI that anticipates the unknown. Subhankar also pointed to machine learning models that can now assess stress and situations that human experts would rarely think of doing on their own, as well as showing early signs in many practices before they appear in a quarterly report.He said that this change changes the work of the financial planner as it changes the work of the engineer: less time collecting numbers, more time thinking about what the numbers mean.“The value is not in the model that produces the answer. The value is in the bank knowing which question was the right one to ask.”Dependence, leadership, and model limitationsSubhankar was careful not to portray the implementation of AI in banking as a technical problem. He used another part of the knowledge of the authority – the need to explain in the decisions of credit and risk, the weight of the management of the financial management, and the high cost of sending models that cannot calculate themselves. In an area where one wrong brand can destroy customer trust and ensure that they feel the same way, he said that proper deployment is as important as possible.This emphasis ties in with a larger argument being made in Panda’s recent work: that trust engineering and AI adoption are not separate conversations. His writing on AI-driven testing in business implementation has provided an important point in the software industry – that as systems grow more autonomous, the discipline to validate them must grow faster, or the speed of AI promises to be a risk instead of an opportunity. Applied to banking, the same logic holds: an AI-driven planning system is just as reliable as testing and leadership built around it.A call for organizational patiencePanda closed by warning against AI transformation in banking as one project with a deadline. He described it instead as a stand-alone capability that needs to be paid for, staffed, and continually reviewed — much closer to how organizations deal with risk management than they do with software releases.“The banks that get this in five years are the ones that see it as infrastructure now, not the ones that are waiting to buy the final product.”ICSCCT 2026 attracted researchers and experts from computing, sustainability, and applied technology for a two-day conference at the University of Malta, which is to be published through Springer.