SHM Converge Scientific Abstract ePoster Gallery-Working 9 to A.I. Giving Clinicians the Mic: Clinician Perspectives From a Multi-Vendor Pilot of Ambient Artificial Intelligence (Ai) Scribes

Working 9 to A.I. Giving Clinicians the Mic: Clinician Perspectives From a Multi-Vendor Pilot of Ambient Artificial Intelligence (Ai) Scribes

Back to course

Pdf Summary

Clinicians across a single academic center were surveyed about AI ambient scribes to understand which features matter most when adopting these tools institutionally. The mixed-methods survey included 69 respondents and asked about current use, preferred features, and concerns.<br /><br />The most commonly used AI scribe was Ambiance, followed by Augmedex. Across responses, clinicians most often said the most helpful feature would be accurate documentation. Other highly valued features included pre-charting the history of present illness, importing clinical data into notes, generating after-visit summaries, and helping place orders.<br /><br />Open-ended comments reinforced that functionality is the main concern. Clinicians emphasized the need for tools that fit naturally into workflow, improve efficiency, and reduce the documentation burden. Key priorities included accuracy, customizability, and strong integration with the electronic medical record (EMR). Some respondents also stressed the importance of vendors continuously improving their products based on physician feedback.<br /><br />The study also highlighted dissatisfaction with frequent transitions between scribe platforms and inconsistent performance across vendors. These shifts were described as disruptive and discouraging, suggesting that stability and reliability are important for frontline adoption.<br /><br />Overall, the findings suggest that clinicians want AI scribes that do more than transcribe—they should reliably support core documentation tasks and integrate smoothly into daily practice. Future research should use larger, more diverse samples and longer-term or experimental designs to better evaluate implementation outcomes.

Asset Subtitle

Pauldeep Mann

Meta Tag

Author List Alan Shindel, James D. Harrison, Kirsten N. Kangelaris, Margaret C. Fang, Pauldeep Mann

Category Research

Concept AI Scribe

Concept Accurate Documentation

Concept Workflow Fit

Concept Clinical Documentation

Concept History of Present Illness

Distinguished Non-Finalist

Presenter Organization University of California, San Francisco

Presenting Author Pauldeep Mann

Track Technology in Hospital Medicine

Keywords

AI ambient scribes

clinical documentation

electronic medical record

mixed-methods survey

physician workflow

documentation accuracy

pre-charting

after-visit summaries

order placement

healthcare technology adoption

AI Scribe

Accurate Documentation

Workflow Fit

Clinical Documentation

History of Present Illness