Implementation and early outcomes of AI-powered ambient documentation in pediatric otolaryngology practice

Electronic health record (EHR) documentation contributes to clinician burnout, yet evidence on ambient artificial intelligence (AI) scribes in pediatric surgical subspecialties is limited. We evaluated adoption and documentation efficiency after implementing an ambient AI scribe in pediatric otolaryngology.

In this retrospective quality-improvement study at a tertiary academic pediatric otolaryngology practice, an ambient AI scribe was introduced in March 2025. We analyzed initial rollout and a current-state window (January-July 2026), after early effects had stabilized, comprising 16 clinicians (12 physicians, 4 advanced practice providers [APPs]). Adoption and documentation time were derived from EHR efficiency dashboards for AI-assisted versus manual encounters. Utilization was compared by visit type and clinician role using chi-square tests and Cochran-Mantel-Haenszel analysis stratified by clinician.

Practice-wide utilization rose from 17.6% of 1690 encounters at rollout to 39.2% the following quarter, and remained 39.9% of 11,565 encounters 10 months later. Adoption was bimodal (6 clinicians ≥50% of encounters; 5 < 10%). Utilization was higher among physicians than APPs (47.7% vs 10.9%; P < 0.001) and in new-patient than established visits (47.0% vs 34.7%; P < 0.001), persisting in clinician-stratified analysis (odds ratio 2.56; P < 0.001). All 16 clinicians remained in the same utilization tier throughout. Median note-editing time was lower with AI assistance (∼4.0 vs 11.8 min), while median note length was similar (∼4000 characters).

Ambient AI scribe implementation was associated with rapid adoption sustained at both practice and individual clinician levels, and with reduced documentation time without increased note length, with particular value for clinicians lacking documentation support.