How might we scale ambient AI at Warp Speed?! This is how. physician exec 14
Good but not great at baseline.
98 Pilot users. 62% retention. Huge waiting list.
Pretty good NPS. 81 Fall 2024. Started at 13.
For top quartile users (caveat). Panama time reduced 50%. 25% decrease time in notes per appt. And same day chart closure up by 10%.
Great quotes.
Really? Go live approach to 5000 in 3 easy steps and only use compatible note template.
Self service materials. 2 minute how-to video. 90% providers self service training. NO MANDATORY TRAINING. 10% used twice a week webinars. no use of kiosks. No use of drop in training.
Smartphrase .AINOTE
Specialties: Late adopter. In person demo: How to get started in 5 minutes to get providers interested.
100-200 per week even with no in person training. Self service.
This was a burnout strategy but also helped with retention. All perceptions on survey.
First time that satisfaction moved significantly.
Press Ganey patient satisfaction improved. Bottom 50% provider performers had much bigger gains.
Actual data: 1 more patient per day with AI scribe. 5% increase in charges.
Unmeasurable value.
Excellent outcome for self service training. FYI vendor was Microsoft DAX.
Q/A. Workflow champion in each specialty. Important.
Problem only one provider per encounter. caution to give it to the app or physician whom ever is doing majority of documentation.
Hard to predict who is going to love it.
Anticipate this will get better quickly.
Support team 1:1 for new provider onboarding.
Least successful in ophthalmology. General dermatology is working, surprisingly. Every specialty there is at least a few high volume users. More templated visits are least useful.
Accuracy and editing. Not quantified.
No increase of patients requesting medical records changes.
Big bang: easier technically. ‘It just works.’ Users helped each other. Adoption rate: target 600 users in 6 months. Reached over 880. out of 5300 providers? For an enterprise license. (Unclear)
Thanks to Vanderbilt!
