EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Med.ai ASK: an agentic system for biomedical question answering

Nhung T H Nguyen, Dmytro S Lituiev, Zhimin Liu, Aditya Kashyap, Garrett Jenkinson, Kevin Kuhl, Christopher Corrado, Naisargi Manishkumar Patel, Kirti Snigdha, Sirwe Saeedi, David Smith, Nicholas Baro und 1 weitere

Journal of the American Medical Informatics Association · 01.06.2026 · 33 (6) · 1134-1145

DOI: 10.1093/jamia/ocag038 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of the American Medical Informatics Association, online erschienen am 30.03.2026. Verfasst von Nhung T H Nguyen und 12 weiteren Autor:innen. Beteiligte Einrichtung: Data Science and Digital Health, Innovative Medicine, Johnson & Johnson, Titusville, NJ, United States. Hauptthemen laut MEDLINE: Information Storage and Retrieval, Artificial Intelligence, Biomedical Research. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

Intelligent agent-driven research co-pilots, leveraging advances in generative AI, are transforming how scientists access biomedical knowledge. This paper presents Med.ai ASK, an agentic question-answering system designed to address biomedical inquiries through dynamic retrieval augmentation and tool-driven reasoning. We aim to develop a system capable of parsing the nuance in biomedical scientists’ research questions to provide reliable, grounded responses that are more accurate than other generative AI solutions. Materials and Methods We adopt the ReAct framework’s tool-calling architecture and leverage atomic reasoning from Self-Discover to build Med.ai ASK. It selectively queries multiple biomedical knowledge bases and employs map-reduce tools for vector database retrieval, alongside external API and NER tool integration. We ingested 44 million biomedical documents from diverse sources. The agent is evaluated on a range of biomedical question-answering datasets.

Results

Human evaluation on an internal dataset shows strong performance and stability. Ratings from a large language model are aligned with human assessments, supporting its use in further experiments. Automatic evaluations indicate superior performance in long-form answers regarding accuracy, faithfulness, factuality, and reduced hallucinations. For short-form and multiple-choice answers, performance is competitive with state-of-the-art systems. The agent’s detailed answers are more interpretable than other systems attributed to its agentic design. The agent effectively selects tools based on question type and is deployed in a production-level chat platform with over 1600 users and 25 000 answered questions.

Conclusion

Med.ai ASK dynamically orchestrates biomedical information retrieval tools to deliver robust interpretative, accurate, and factual answers, which is crucial in the biomedical domain.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Information Storage and RetrievalArtificial IntelligenceBiomedical ResearchHumansGenerative Artificial IntelligenceIntelligent SystemsLarge Language ModelsNatural Language Processing

Stichwörter der Autor:innen

agentic AIbiomedical question answeringlarge language modelreasoningretrieval-augmented generation

Autor:innen und Institutionen

  • Data Science and Digital Health, Innovative Medicine, Johnson & Johnson, Titusville, NJ, United States.
  • Architecture and Engineering, JJT; Johnson & Johnson, Raritan, NJ, United States.
  • Discovery, Product Development & Supply, Innovative Medicine; Johnson & Johnson, San Diego, CA, United States.

Zitieren

APA

Nguyen, N. T. H., Lituiev, D. S., Liu, Z., Kashyap, A., Jenkinson, G., Kuhl, K., Corrado, C., Patel, N. M., Snigdha, K., Saeedi, S., Smith, D., Baro, N., & Schultz, T. (2026). Med.ai ASK: an agentic system for biomedical question answering. Journal of the American Medical Informatics Association, 33(6), 1134-1145. https://doi.org/10.1093/jamia/ocag038

Vancouver

Nguyen NTH, Lituiev DS, Liu Z, Kashyap A, Jenkinson G, Kuhl K, et al. Med.ai ASK: an agentic system for biomedical question answering. Journal of the American Medical Informatics Association. 2026;33(6):1134-1145. doi:10.1093/jamia/ocag038

BibTeX
@article{Nguyen2026,
  author = {Nguyen, Nhung T H and Lituiev, Dmytro S and Liu, Zhimin and Kashyap, Aditya and Jenkinson, Garrett and Kuhl, Kevin and Corrado, Christopher and Patel, Naisargi Manishkumar and Snigdha, Kirti and Saeedi, Sirwe and Smith, David and Baro, Nicholas and Schultz, Timothy},
  title = {Med.ai ASK: an agentic system for biomedical question answering},
  journal = {Journal of the American Medical Informatics Association},
  year = {2026},
  volume = {33},
  number = {6},
  pages = {1134--1145},
  doi = {10.1093/jamia/ocag038}
}
Kostenlos starten →