About OpenSupport
Generative AI a regulated team can approve.
Health questions arrive at every hour, and the answers people find on their own are rarely sourced. We build an assistant whose every answer can be traced, checked and governed — so the teams responsible for what patients are told can put it in front of them.
What we build
An embeddable AI assistant for medical information and patient support programs. It answers questions about conditions, treatment, access and affordability from a graded corpus of 16,500+ sources — and from a deployment's own approved content — cites its sources, and defers case questions to the people and channels who can answer them.
How we approach it
Retrieval before generation, so an answer is assembled from sources rather than recalled. Guardrails that run on every answer and that no deployment's configuration can switch off. A privacy posture built not to hold personal health information: identifiers are redacted before an answer is generated, and every answer is checked for personal health information once it is. And corpus figures on this site that are generated from production rather than written — the editorial policy sets out how.
What we hold ourselves to
Every claim traceable to its source. Retrieval and ranking identical for every deployment, so no one can buy a higher position in an answer. Health information, not medical advice: no diagnosis, no prescribing, and a clear route back to a healthcare professional.
Who we work with
Medical affairs, medical information and patient-services teams who want generative AI they can defend in review. Request a demo, or try the assistant against your own questions.
Give patients a better answer before they give up searching.
See how OpenSupport supports disease education, patient services, market access and enterprise patient engagement.