There is a deluge of published information in healthcare. Considering just the tens of millions of abstracts publicly available in PubMed, it’s already extremely difficult and time-consuming to find the evidence you need to make a decision (let alone stay current on a healthcare topic). In this talk, we will describe a commercial system that uses AI to automatically surface relevant results published in the medical literature. Rather than having to read every piece of content, users can search and filter the articles as they might in a database or spreadsheet. And by using AI, we can do so at a massive scale across diverse content in healthcare including clinical outcomes, epidemiology, economic data, and even humanistic data (e.g., patients’ health-related quality of life). In addition to describing our NLP, we will demonstrate the system itself and speak to how it all works. By treating the medical literature as a ” database of evidence” rather than articles to read, we can unlock huge efficiencies (sometimes 59X-faster than human efforts!), freeing humans to determine what treatment options exist for different diseases, and what evidence exists for their safety and effectiveness among patient populations. The platform can identify data gaps that need to be filled with additional research, while also even automating comparative effectiveness research to determine which treatments may have more evidence as to their safety and effectiveness for given patient populations.