By: ICN Bureau
Last updated : March 02, 2026 12:31 pm
The technology solves complex chemical structure identification in polymer/organic research and generates comprehensive reports, including compound properties, uses, and manufacturing methods
Mitsui Chemicals has developed a proprietary AI agent designed to autonomously analyze, investigate, and organize compound information from scientific literature and research reports based on chemical structural formulas. Following successful initial verification, which indicated a minimum 80% reduction in time spent on literature reviews, the company has initiated an in-house pilot trial.
Mitsui Chemicals plans to conclude this trial before the end of fiscal 2025, with full-scale implementation scheduled for fiscal 2026 to enhance R&D productivity.
The investigation of compound information found in patent literature and academic papers is an essential part of chemical R&D. However, in the case of polymers and organic compounds with complex structures, it is difficult to accurately identify a compound’s name from its structural formula alone.
Until now, researchers have had to expend enormous time and effort in manually checking vast amounts of literature, piece by piece, to determine the correspondence between chemical structures and compound names. By combining the use of the newly developed AI agent with the chemical expertise of the company’s researchers and engineers, Mitsui Chemicals has achieved the automation of advanced literature searches in a manner that was not possible using conventional methods.
This AI agent enables AI to read chemical structural formulas – something that had previously been unfeasible – and thereby allows AI agents to employ image and text information from literature in an integrated manner to autonomously identify and extract information about compounds. In addition to identifying compound names, the AI agent can simultaneously obtain related information recorded in the literature, including uses, physical properties, manufacturing methods and experimental conditions.
Furthermore, the agent generates more comprehensive compound information after consulting chemical databases and web search information if needed. It is also equipped with a report generation feature tailored to researchers’ needs, which organizes the survey results into a format that supports subsequent analysis and decision-making.
Initial validation confirmed that literature reviews conducted using this AI agent reduced the time spent by researchers on investigation by at least 80 percent. Literature reviews that previously took about a month were shortened to a day or so, allowing researchers to spend more time focusing on creative R&D activities.
Mitsui Chemicals will continue utilizing cutting-edge digital technologies to address social challenges with the power of chemistry, and to contribute to a sustainable society through the creation of diverse value.