Abstractive Health

Abstractive Health is a New York-based clinical AI company that retrieves a patient's longitudinal medical record from national health information exchange networks and turns it into an AI-generated summary for clinicians ahead of a visit. Its stated mission is to make the medical record reviewable before treatment decisions are made. The product spans three core engines: record retrieval, AI clinical and medical record summarization, and retrieval-augmented clinical question answering, with an additional AI clinical decision support alert capability.

Record retrieval runs through a Carequality integration and performs geospatial retrieval around a patient's current and prior addresses, returning a unified chart rather than disconnected encounters. Incoming data formats include CCDAs, FHIR R4 resources, PDFs and image-based documents, which are normalized into a single JSON structure while original CCD and FHIR files remain downloadable. The summarization engine uses a section-to-meta pipeline to surface diagnoses, medications, allergies, hospitalizations and key events across institutions, linking each sentence back to its source note for sentence-level provenance. The RAG engine combines the patient record with medical references to return cited answers to clinical questions.

The company describes itself as a physician AI assistant deployed as a SMART on FHIR application, integrating with national HIE networks including Carequality and CommonWell and with major EHR systems. It states HIPAA compliance and SOC-2 Type 2 certification.

Founding story

Abstractive Health was founded by Vince Hartman (Co-founder & CEO), Giordana Pulpo (Co-founder & Head of Design) and Ritika Poddar (Co-founder & CTO), who built the company out of Cornell Tech, where Hartman completed a Master of Science in Computer Science with a health tech concentration, Poddar a Master of Science with a Connective Media concentration, and Pulpo participated in Product Studio and Startup Studio.

Business model

Abstractive Health sells software subscriptions to individual clinicians and organizations, plus API access and enterprise EHR integrations. A free clinician tier includes unlimited AI question answering, five patient searches with AI summaries per month and five PDF uploads for summarization per month. A standalone web app plan is priced at $99 per user per month for up to five users with unlimited patient searches, summaries and scan summarizations, export to PDF/FHIR/JSON, batch schedule uploads and frontend EHR integration. An Organization plan, required above five users, is custom-priced and adds API access, backend EHR integration, implementation support for EHR, SFTP and API workflows, enterprise contracting and security review, and priority support.

Per-user monthly subscriptions (free tier and $99/user/month standalone plan) plus custom-priced organization contracts covering full EHR and API integration.

Traction

The website cites clinician customers at Greater Good Health, Sally Health and Moon Medicine, published work in JAMA Network Open with Weill Cornell Medicine, case studies on care transitions and VA-ready summaries of community care records, a reported ~39% improvement in comprehension accuracy from its AI clinical summaries, and a stated ~30% reduction in average visit prep time. As of a Forbes profile last updated January 2024, the company was running a study with NewYork-Presbyterian with a pilot expected to go live within months.

Latest developments

Recent company blog posts cover sentence matching in AI clinical summarization (November 2025), record retrieval and summarization standards (August 2025) and retrieval-augmented generation in healthcare (July 2025). Current offerings include a developer API with sandbox access and documentation, CME credits, and case studies on care transitions and converting community care records into VA-ready AI summaries.