Patsnap Launches AI Agents to Extract SAR Data and Identify Lead Compounds From Small Molecule Patents

Patsnap today announced Lead Compound Analyzer and SAR Extractor, two AI agents in Patsnap Eureka Life Sciences that help drug discovery teams identify likely lead compounds and extract structure-activity data from small molecule patents. By reducing manual extraction, the agents help medicinal chemistry, competitive intelligence and computational teams assess competitor compounds, compare chemical series and inform further research. Learn more here.

AI is giving drug discovery teams new ways to analyze compounds and guide molecular design, but those workflows depend on the experimental data available to them. Patents are often an early public source of new compounds, potency data and structure-activity relationships, yet that evidence remains difficult to reuse.

Compound structures appear as images, activity values sit in tables and the legends that explain them can be pages apart. Connecting those details by hand takes time. The new agents help turn that scattered evidence into structured, source-referenced data that scientists can review and use in computational research.

“Patents disclose valuable evidence about which compounds were made, how they were tested and what the results showed. Making that evidence usable still takes substantial scientific effort,” said Mark Davies, Head of Product, Life Sciences, at Patsnap. “These agents bring structures and activity data together, with references scientists can check, so teams can spend more time assessing the chemistry and deciding what to investigate next.”

The two agents address complementary parts of this challenge:

Lead Compound Analyzer identifies up to 10 compounds most likely to be the lead in a small molecule patent, grouped by the strength of the biological evidence disclosed. Supporting rationale and source references help scientists assess the candidates and determine where further investigation is needed.

SAR Extractor processes up to 10 patents per task, connecting compound structures with reported activity values, targets and experimental context in a structured table. Patsnap’s optical chemical structure recognition (OCSR) model achieved 97% precision in testing across more than 420 randomly selected English-language patents, covering over 8,800 structures and 5,100 complete molecules. Separate confidence scores for structure and activity, alongside source references, help scientists review the extracted data and validate findings.

Used together, the agents help teams assess likely lead compounds alongside the wider chemical series disclosed in patents, supporting competitor assessment and structure-activity analysis. They complement Patsnap Chemical and Synapse within Patsnap’s broader portfolio of chemistry, target and drug development intelligence.

Davies will present “Patents In, SAR Out: Grounding Agents in Chemistry” at the AI Drug Discovery & Development Summit in Boston on 29 October.

Both agents are available now. Drug discovery, competitive intelligence and computational chemistry teams can request a demonstration using patents relevant to their research here.

About Patsnap

Founded in Singapore in 2007, Patsnap is a leading global innovation intelligence company. Leveraging domain-specific AI agents and robust analytics tools, Patsnap empowers IP, R&D and Life Sciences teams to unlock insights, accelerate discoveries, and protect breakthroughs. Trusted by nearly 20,000 enterprises, law firms, and research institutions in over 50 countries, including one-third of the world’s largest R&D spenders, Patsnap transforms how organizations innovate and make critical decisions. Learn more at www.patsnap.com.

View source version on businesswire.com: https://www.businesswire.com/news/home/20261027520635/en/