Banking organizations face increasing pressure to achieve digital transformation while maintaining operational stability, regulatory compliance, and service reliability. Although previous studies have explored enterprise architecture and TOGAF-based digital transformation, limited research has translated these approaches into technology architecture artifacts that simultaneously support governance and AI-driven innovation in banking. This study proposes a TOGAF-based ambidextrous technology architecture for BankCo’s lending process using design science research methodology (DSRM). Unlike conventional TOGAF-based approaches, the proposed architecture integrates traditional and agile technology perspectives through governance mechanisms, technology modernization strategies, and architecture decision models to support continuous technology evolution. The resulting architecture captures technology principles, standards, governance readiness assessment, modernization strategy, an iterative implementation roadmap, and an ambidextrous decision model for AI-enabled lending transformation. The stakeholder-based architecture validation shows that the proposed architecture is feasible for supporting operational stability and enabling the controlled deployment of AI-driven lending capabilities. This research operationalizes ambidextrous enterprise architecture within the technology architecture domain. It provides banking organizations with a structured approach for introducing AI-driven lending capabilities while maintaining governance, regulatory compliance, and operational stability.