「特許分析」機能と「新規用途探索」機能については、事業部や研究開発部門の協力を得て行った実験において、80%の業務時間削減効果が確認されました。特に、構造式や表形式データを含む技術文書の解析において、従来の手法と比較して大幅な効率化が期待できます。また、「営業提案」機能は、文章データのみならず、表形式のデータや化合物の構造式からも情報を抽出・解析することが可能で、化学分野特有の複雑な情報処理ニーズに対応しています。Mitsui Chemicals, Inc. (Headquarters: Chuo-ku, Tokyo, President: Osamu Hashimoto) has developed a platform equipped with a uniquely developed generative AI chat to meet the highly specialized business needs unique to the field of Chemicals, such as interpreting experimental results in tables and understanding chemical structure formulas. A proof of concept will be completed through use in business divisions and research and development departments within the 2024 fiscal year, and full-scale operations within the company will begin from the 2025 fiscal year.
This initiative aims to promote digital transformation (DX) in the business area and to create ideas for new product development and expand product top-line (sales) and Market Overview Share. By combining the specialized knowledge of Chemicals held by our researchers and engineers with generative AI, it will lead to new values that could not have been found by traditional methods.
The platform developed this time is equipped with three functions. First, it has a patent analysis function that responds to user inquiries from a vast amount of patent information, which can be utilized for patent investigations, issue extraction, and technology trend research. Second, it incorporates a new application exploration function that derives essential information for exploring new applications from large-scale patent information, supporting the discovery of new applications and the planning of potential applications. Third, it implements a sales support function that extracts insights related to sales activities from internal information and web information, which can be utilized for market research, prospect discovery, and competitive analysis.
For the "patent analysis" and "new application exploration" functions, it has been confirmed through experiments conducted with the cooperation of business divisions and research and development departments that there is an 80% reduction in work time. In particular, significant efficiency improvements are expected in the analysis of technical documents containing structural formulas and tabular data compared to traditional methods. Additionally, the "sales proposal" function can extract and analyze information not only from text data but also from tabular data and chemical structure formulas, addressing the complex information processing needs unique to the Chemicals field.
The insights gained through this initiative will actively promote collaboration with academia and other companies in a bid to contribute to the development of the entire Chemicals industry.
That's all.
