Assists in using clinical guidelines to make decision making quicker for healthcare professionals
Medwise AI automates healthcare professionals’ search for clinical guidelines. Its question-answering model informs decision making at the point of care. This uses proprietary agentic AI to search and extract snippets of text from evidence-based best practice guidelines- not only from public internet sources but also local content that may be uniquely owned and stored on an organisation’s intranet. This is non-generative AI, with no risk of hallucinations, for safety and quality in patient care.
User asks a question, model returns answer from text-based documents including PDF as well as HTML
Answer provided is a snippet from a cited source showing the location in the document
The model is built on trained and finetuned open source and close source models. LLM powers an agentic AI workflow to accomplish tasks, such that the search simulates a clinician looking for answers and therefore only selects and displays snippets or documents. This ensures safety and reliability for healthcare questions because 1) only pre-approved sources are indexed and used for answers, and 2) outputs are snippets and thus free of hallucinations. Every action is 100% traceable and auditable, with 0% risk of LLM-generated hallucinations. No new language is generated: all text responses and answers are a parsed quote from a cited source.
An independent evaluation conducted by Health Innovation East found an estimated average time saving of 2.6 minutes per 10 minute consultation.
An independent health economic evaluation completed by Health Innovation East showed up to £30M productivity savings for the NHS.
Respondents to a survey conducted by Explain Market Research Ltd reported high satisfaction levels (average score of 7.5 out of 10) and confidence in the information gathered (91%). They also stated that Medwise.ai allowed them to access clinically relevant information faster (82%).
