EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Agentic AI for Banking: A Predictive Framework for Credit Risk Assessment, Fraud Detection, and Customer Intelligence

AashiKrishna Ravichandran

International Journal of Computational and Experimental Science and Engineering · 30.05.2025 · 11 (4)

DOI: 10.22399/ijcesen.5372 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computational and Experimental Science and Engineering, online erschienen am 30.05.2025. Verfasst von AashiKrishna Ravichandran. Die Arbeit nennt 11 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The growing trend of more financial frauds, chances of credit defaults, and personalization of customer information has exposed the shortcomings of conventional single model AI systems used in banking. This paper proposes, designs, and analyzes the Agentic AI Predictive Framework for banking through the integration of three agentic autonomous agents – a Credit Risk Agent, a Fraud Detection Agent, and a Customer Intelligence Agent – through the CrewAI-based multi-agent system architecture. The agentic agents will be designed based on XGBoost and reinforcement learning algorithms, and will use public domain data such as UCI Credit Default Data Set and IEEE CIS Fraud Detection Dataset. A simulation experiment is conducted to compare the efficiency of the proposed framework with that of a baseline framework using logistic regression, random forest, and deep learning models through the aforementioned datasets. An explainable framework through SHAP for all the agents will make decision output explanation feasible. From these results, it can be said that MAS is able to predict credit risk at 92.6%, detect fraud at 94.7%, and predict customer engagement at 89.4% level of accuracy, outperforming all baselines consistently. Cooperation among the agents led to 11.3 percentage point improvement in fraud detection accuracy from individual agent performances, showing the usefulness of combined intelligence in the agentic approach. Regulatory compliance has been attained at 88.6% in accordance with Basel III and EU AI Act regulatory compliance requirements. It can be said that agentic AI has brought about revolutionary changes in bank risks’ management.

Zitieren

APA

Ravichandran, A. (2025). Agentic AI for Banking: A Predictive Framework for Credit Risk Assessment, Fraud Detection, and Customer Intelligence. International Journal of Computational and Experimental Science and Engineering, 11(4). https://doi.org/10.22399/ijcesen.5372

Vancouver

Ravichandran A. Agentic AI for Banking: A Predictive Framework for Credit Risk Assessment, Fraud Detection, and Customer Intelligence. International Journal of Computational and Experimental Science and Engineering. 2025;11(4). doi:10.22399/ijcesen.5372

BibTeX
@article{Ravichandran2025,
  author = {Ravichandran, AashiKrishna},
  title = {Agentic AI for Banking: A Predictive Framework for Credit Risk Assessment, Fraud Detection, and Customer Intelligence},
  journal = {International Journal of Computational and Experimental Science and Engineering},
  year = {2025},
  volume = {11},
  number = {4},
  doi = {10.22399/ijcesen.5372}
}
Kostenlos starten →