February 10, 2025

  • Updated
Download Icon Download

Search Enhancement: Reaction Natural Language Search Improvements

Text searching reactions in CAS SciFinder has been enhanced with AI-driven query interpretation. AI, trained with CAS data, is being leveraged in CAS SciFinder for the first time to provide precise, accurate query interpretation resulting in focused and highly relevant results. Note: The AI is trained with CAS data and is not generating data; all results presented are CAS content.

SF-2025-02-10-1.png

This improved natural language query support will include supporting queries containing substance names, CAS RNs, molecular formulas, substance classes, and functional groups to identify reaction participants.

Search Enhancement: Expanded Reaction Role Recognition

Searching reactions in CAS SciFinder has been enhanced with expanded reaction role recognition. When a reacting participant is identified in the query, the query is expanded to include the participant in reactant or reagent role.

SF-2025-02-10-2.png

Search Enhancement: Expanded Reference Document Identifier Recognition

Searching for references using document identifiers has been enhanced with improved query recognition. When a document identifier is detected, SciFinder determines a reference intent. This search enhancement has the benefit of recognizing more variations in document identifiers; for example, patent numbers can be searched including kind codes, with and without spaces in the query.

SF-2025-02-10-3.png

Workflow Enhancement: Improved XLSX Export Performance

Reference result exports in XLSX format have been updated to improve performance and reduce extraneous data. When users export references in XLSX format, the export menu option will no longer automatically include all available content. Users may now select which data they want to include in the exported content.

SF-2025-02-10-4.png