Sakana AI and Mitsubishi UFJ Financial Group (MUFG) have officially entered the verification phase for their "AI Lending Expert," a specialized tool designed to revolutionize bank loan operations. This ambitious project seeks to redefine professional workflows by employing Agentic AI capable of complex logical construction and autonomous reasoning. Rather than merely summarizing data, the system aims to automate the intricate decision-making processes required in high-level finance.

The project's core strength lies in applying the methodologies behind Sakana AI's "AI Scientist" and "ALE-Agent" to real-world banking tasks. The AI autonomously organizes and executes the collection and analysis of vast datasets, moving away from a model that requires constant human oversight for every sub-task. Crucially, the system captures the "tacit knowledge" of veteran employees, structuring it into a format that the AI can leverage to ensure outputs remain consistent with the bank's specific standards and institutional logic.

An extraordinary improvement cycle drove the development process, where the AI categorized and analyzed over 1,500 detailed feedback points from bank staff to automatically refine its own prompts and operating rules. This automated precision enhancement allowed the system to quickly mature from a novice-level assistant into a sophisticated "buddy" for professional bankers. Ultimately, the project demonstrates how AI can handle repetitive cognitive loads, freeing human experts to focus on nuanced customer relationships and qualitative assessments that require a human touch. This collaboration could set a new standard for Generative AI in the Japanese financial sector.

Sakana AI and Mitsubishi UFJ Financial Group (MUFG) have officially entered the verification phase for their "AI Lending Expert," a specialized tool designed to revolutionize bank loan operations. This ambitious project seeks to redefine professional workflows by employing Agentic AI capable of complex logical construction and autonomous reasoning. Rather than merely summarizing data, the system aims to automate the intricate decision-making processes required in high-level finance.

The project's core strength lies in applying the methodologies behind Sakana AI's "AI Scientist" and "ALE-Agent" to real-world banking tasks. The AI autonomously organizes and executes the collection and analysis of vast datasets, moving away from a model that requires constant human oversight for every sub-task. Crucially, the system captures the "tacit knowledge" of veteran employees, structuring it into a format that the AI can leverage to ensure outputs remain consistent with the bank's specific standards and institutional logic.

An extraordinary improvement cycle drove the development process, where the AI categorized and analyzed over 1,500 detailed feedback points from bank staff to automatically refine its own prompts and operating rules. This automated precision enhancement allowed the system to quickly mature from a novice-level assistant into a sophisticated "buddy" for professional bankers. Ultimately, the project demonstrates how AI can handle repetitive cognitive loads, freeing human experts to focus on nuanced customer relationships and qualitative assessments that require a human touch. This collaboration could set a new standard for Generative AI in the Japanese financial sector.