Integrating AI as an enterprise architecture level design element for financial institutions

Abstract

Artificial Intelligence (AI) integration in highly regulated financial institutions can present both opportunities and challenges. Already existing EA frameworks and AI Canvases often fail to fully address the unique requirements of AI in the banking sector, such as data governance, regulatory compliance, ethical considerations and cross-functional alignment. This thesis addresses this gap by developing an AI Canvas for Enterprise Architecture, a conceptual framework that integrates governance, risk and compliance elements while coordinating AI initiatives with the traditional EA layers of Business, Data, Application and Technology. The research follows a Design Science Research (DSR) methodology in combination with semi-structured expert interviews with practitioners from financial institutions and consulting firms. It helps to significantly improve the proposed AI Canvas. After that it was assessed by industry professionals. The expert evaluation was performed using completeness, clarity, usability, practical relevance, stakeholder acceptance and decision-support value criteria. The findings show that the AI Canvas offers enterprise architects in financial institutions a useful and understandable tool, as it was highly evaluated by the practitioners. The thesis contributes both academically and practically, by extending EA research with an AI-focused framework and offering a validated tool for EA practitioners to plan an AI integration early in the financial sector.