New natural language capability aims to reduce barriers for researchers and speed up innovation across life sciences

Elsevier has launched Reaxys AI Search, a new natural language tool designed to accelerate chemistry-led research across biopharma, materials science, and polymer development. The AI-powered feature is integrated into the Reaxys database and aims to reduce the time scientists spend navigating complex datasets and technical literature by interpreting questions written in plain language.

Elsevier says Reaxys AI Search is the first natural language tool built into a comprehensive chemistry research database. Rather than relying on keyword strings or structured query logic, users can now enter questions as they would naturally ask them. The system understands synonyms, spelling variations, and scientific terminology, returning relevant results from more than 121 million documents, including patents and peer-reviewed journals.

The tool has been developed in collaboration with hundreds of chemists and is already available to Reaxys users in its early-access version. The platform continues to include its existing structure-based and keyword search features.

“Reaxys AI Search marks a major step in making chemistry data more accessible and actionable,” said Mirit Eldor, managing director of life sciences at Elsevier. “By enabling natural language queries, we aim to lower barriers for researchers across disciplines and experience levels to find the information they need faster and with greater confidence. Launching this early version allows us to deliver immediate value to researchers while gathering feedback to continually refine the solution as part of our commitment to innovation that helps advance human progress.”

The tool is particularly suited to interdisciplinary work in R&D environments where chemistry overlaps with biology, engineering, and pharmaceutical sciences. It builds on earlier AI enhancements to Reaxys, including its predictive retrosynthesis tool, and represents a broader shift toward AI-assisted literature review, study design, and early formulation work.

Elsevier emphasises that Reaxys AI Search was built according to its Responsible AI Principles and Privacy Principles. The company says that user queries are private and not used to train external models, and that search results are always generated from curated, trusted sources.

Additional features, including AI-generated summaries, are planned in future releases.