This study examines the ambivalent effects of agentic artificial intelligence (Agentic AI) on the sustainable financial sector. The analysis shows that these systems not only respond to user needs, but are also capable of proactively making financial decisions, ushering in a new era in banking services (McKinsey & Company, 2025). Agentic AI offers significant efficiency gains, cost reductions and a personalised customer experience (EY, 2023a; Deloitte, 2025), while also enabling greener operations through process optimisation and reduced resource use. At the same time, it carries new ethical, regulatory and technological risks. The opacity of algorithms (“black box”), discriminatory biases and data protection vulnerabilities can undermine user confidence and, if left unchecked, threaten both financial resilience and the broader sustainability of the sector (Körber, 2019; Ryan, 2020; Lendvai & Gosztonyi, 2025). Particular attention should be paid to the erosion of the “inertia dividend”, which signals the gradual disappearance of bank profits derived from customer passivity. Through real-time optimisation by AI agents, customer funds automatically flow into more favourable and often more sustainable financial structures, eliminating banks’ hidden sources of profit (Boston Consulting Group, 2025). This process will lead to a decline in net interest margins and card business revenues, as well as the transformation of loyalty-based customer relationships into performance-driven models that reward transparency and responsibility (Financial Times, 2024; McKinsey & Company, 2025). In the longer term, Agentic AI will create a new business logic in which algorithms stand at the centre of customer relationships. This will result in a more dynamic, transparent, but also more vulnerable competitive market. The concentration of technological dependence, algorithmic homogeneity and regulatory asymmetries pose significant systemic risks, including sustainability risks related to resilience, inclusion and the environmental footprint of digital infrastructures (FSB, 2024; IOSCO, 2025; ECB, 2024b). The study concludes that banks must adapt by developing sustainable and resilient revenue models, while regulators need to establish stricter transparency and supervisory frameworks. Only through such measures can artificial intelligence support efficiency, innovation, and long-term financial stability in a manner that is consistent with environmental, social and governance (ESG) objectives and broader sustainability requirements.
XIII. ÉVFOLYAM 2025. SPECIAL ISSUES 2. 2-7
DOI: 10.24387/CI.SI.2025.2.1