Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

This is some text inside of a div block.
This is some text inside of a div block.

Problem Tech Solves

Tech Brief

Radiologist burnout has been named the biggest threat to radiology by AuntMinnie for three years in a row. US medical imaging volume continues to rise, and radiologist shortages continue to worsen. Radiologists spend 75-80% of their time manually dictating reports using voice recognition; some dictation occurs while reviewing images, but the majority of our time as radiologists is spent dictating reports. The impression is considered the most important part of the report, as it contains conclusions and follow-up recommendations; surveys find that many ordering clinicians only read report impressions. In working with thousands of radiologists, we’ve found that 30-35% of all manually dictated words are in the impression. Macros are commonly usable in the Findings, but the impression requires much more manual dictation. By automatically generating impressions customized to each radiologist’s language and style, we reduce the number of words spoken each day by one-third, addressing radiologist fatigue and burnout, while saving radiologists a median of 1 hour per 9-hour shift. Part of that time savings becomes increased radiologist productivity, which is key for radiology practices inundated by rising study volumes. Rad AI Omni also improves report accuracy: helping the radiologist catch and fix errors they made in the Findings section, ensuring that significant incidental findings are included in the impression, and inserting standardized follow-up recommendations based on national consensus guidelines. The combination of reduced burnout, increased productivity, and improved report accuracy provide ROI unmatched by any other radiology AI product, helping explain our rapid growth and recent awards/recognition.

Tech Differentiators

Rad AI Omni is built specifically by radiologists, for radiologists, and embeds seamlessly into existing workflow as a zero-click solution. A radiologist leader at one of the largest practices in the country noted that “when we deployed [another Best in Class solution], it took our radiologists about a year to get used to the product in their workflow. For Rad AI Omni, it took most radiologists a day.” Rad AI Omni continues to improve rapidly – impression accuracy and language/style customization increases with each new set of models, product speed has increased by 3-4x since launch, application features continue to expand, and we address radiologist feedback/requests as promptly as possible. This helps ensure a highly positive customer experience and raises the competitive barrier to entry. There are currently no competitors for Rad AI Omni, as it is an unusually difficult product to build well. Rad AI has by far the largest dataset of historical reports among startups (over 150 million reports) and continues to improve each radiologist’s language customization based on their edits to impressions, building a proprietary dataset of millions of additional reports every quarter. Its seamless integrations are also currently unique; several vendors are seeking partnerships specifically to benefit from Rad AI’s reporting integrations. Finally, there are direct synergies with our other products; Continuity, our product for tracking and closing the loop on follow-up for incidental findings, generates even stronger ROI for health systems when paired with Omni’s real-time improvement of recommendations language and inclusion of significant incidental findings.


An analysis by Radiology Partners during their initial deployment of Rad AI Omni to over 100 radiologists found a mean increase in radiologist productivity of 11% after deployment (based on RVUs, compared to the baseline period immediately prior). Rad AI’s analysis of radiologist time savings with more than 500 radiologists at 20 radiology practices across the US found average time savings ranging between 10% and 14% at different practices, after deployment of Rad AI Omni (compared to the baseline period immediately prior). Einstein published an abstract at RSNA 2021 assessing the accuracy of CT impressions generated by Rad AI, across a random selection of 120 historical CT reports from Einstein. They found no statistical difference in accuracy between Rad AI-generated impressions and original radiologist impressions, and no statistical difference in radiologist preference between the two. Rad AI performs regular NPS surveys of radiologist users of Rad AI Omni. This year so far, Omni has an NPS score of 62 from its radiologist users. For quick comparison, Apple is reported to have an average NPS score of 61 over the past two years. This indicates strong radiologist advocates and champions among our partner groups. It is the vocal support of radiologists, practice leaders and many others that has helped make our recent recognition and awards possible.

Why Us