Winsystem Achieves 10x Development Productivity Gain with Amazon Bedrock
Japanese software company Winsystem utilized Amazon Bedrock and an AI coding agent to significantly improve internal tool development efficiency.
Tokyo – Winsystem, a software development company, has dramatically enhanced its development productivity by integrating Amazon Web Services (AWS) Bedrock, a service for building and scaling generative AI applications, with an AI coding agent named Cline.
Winsystem, known for developing systems for the travel industry and its corporate travel management product "Travel Studio," faced challenges with outdated internal management tools. The need to improve the efficiency of internal operations and digitize them into web applications was pressing, but the company also had a strategic imperative to focus development resources on client projects.
The company addressed this by combining the Cline AI coding agent with Amazon Bedrock, specifically utilizing the Claude 3.7 Sonnet model. This AI-driven development environment enabled the creation of a web application version of their internal management tool in just one business day. Previously, a comparable task would have required an estimated ten business days for technical investigation and development, representing a tenfold increase in productivity.
"By utilizing an AI coding agent, development productivity increases dramatically," stated Hayato Takahashi, an engineer from Winsystem's systems department. "Using Amazon Bedrock as our LLM provider allows us to develop while maintaining data privacy and governance."
The AI-assisted development has also been applied to accelerate sales proposals by creating prototypes within hours, improve the quality of existing systems through code refactoring, and expedite the multi-language conversion of existing websites to facilitate global expansion.
Winsystem plans to further expand its use of AI-driven coding to automate more complex yet repetitive tasks and optimize its selection of AI tools for cost-effectiveness. Additionally, they intend to establish documentation outlining project-specific constraints and the company's coding standards for the AI agent to reference.