Sakana AI launched its Applied Team in early 2025 to implement generative AI, inspired by collective intelligence in nature, into social infrastructure fields like finance and defense. Software engineers Shota Sakai and Katsuhiro Honda are currently navigating the complexities of integrating AI agents into real-world business environments.

Sakai is overseeing the development of an AI agent designed to assist with bank loan operations. The project aims to automate tasks such as information gathering, financial analysis, and drafting loan proposals, allowing bank employees to focus on critical decision-making. The core technical challenge lies in defining and ensuring 'operational quality' while navigating the strict security requirements and legacy system constraints inherent in the financial industry. This requires a comprehensive full-stack design that encompasses UI/UX, error recovery mechanisms, and audit logging.

The engineering team operates through close collaboration between software and solution engineers. In enterprise sectors like finance, success depends not only on implementation capabilities but also on integrating systems into client environments and establishing a cycle where field feedback informs platform improvements. The company has moved beyond using AI merely as a development aid, now focusing on restructuring organizational workflows with AI as a fundamental premise. This shift has expanded the scope of individual engineer responsibilities from infrastructure maintenance to evaluating product value.

Honda emphasizes that as AI-assisted coding changes the nature of software engineering, the goal is to build a development organization that delivers grounded, tangible value to society. Sakana AI maintains an approach that avoids both overestimating and underestimating AI, instead utilizing rigorous governance and validation environments to ensure humans remain focused on final risk assessment. The company plans to codify the insights gained from these projects to establish a de facto standard for AI adoption among Japanese enterprises.

Sakana AI launched its Applied Team in early 2025 to implement generative AI, inspired by collective intelligence in nature, into social infrastructure fields like finance and defense. Software engineers Shota Sakai and Katsuhiro Honda are currently navigating the complexities of integrating AI agents into real-world business environments.

Sakai is overseeing the development of an AI agent designed to assist with bank loan operations. The project aims to automate tasks such as information gathering, financial analysis, and drafting loan proposals, allowing bank employees to focus on critical decision-making. The core technical challenge lies in defining and ensuring 'operational quality' while navigating the strict security requirements and legacy system constraints inherent in the financial industry. This requires a comprehensive full-stack design that encompasses UI/UX, error recovery mechanisms, and audit logging.

The engineering team operates through close collaboration between software and solution engineers. In enterprise sectors like finance, success depends not only on implementation capabilities but also on integrating systems into client environments and establishing a cycle where field feedback informs platform improvements. The company has moved beyond using AI merely as a development aid, now focusing on restructuring organizational workflows with AI as a fundamental premise. This shift has expanded the scope of individual engineer responsibilities from infrastructure maintenance to evaluating product value.

Honda emphasizes that as AI-assisted coding changes the nature of software engineering, the goal is to build a development organization that delivers grounded, tangible value to society. Sakana AI maintains an approach that avoids both overestimating and underestimating AI, instead utilizing rigorous governance and validation environments to ensure humans remain focused on final risk assessment. The company plans to codify the insights gained from these projects to establish a de facto standard for AI adoption among Japanese enterprises.