Ambient AI scribe in pediatric outpatient visits: documentation burden, quality, and scale-up limits.

INTRODUCTION: Artificial intelligence (AI) medical scribes are an emerging technology intended to reduce the documentation burden. In pediatrics, evidence on implementation in the Spanish context is very limited. OBJECTIVE: To evaluate the implementation of an AI medical scribe in pediatric outpatient clinics, examining its effects on clinician-patient interaction, documentation quality, and barriers to adoption. METHODS: Prospective observational proof-of-concept study (February-October 2025) conducted at a tertiary pediatric hospital. Forty physicians from three subspecialties participated, and 197 visits were documented. The evaluation combined direct observations (n = 30), surveys of clinicians (n = 18) and patients (n = 20), and a quality assessment using a rubric based on the mPDQI-9. RESULTS: With the AI scribe, eye contact increased (median 95.7% vs. 78.6%; p < 0.001), with no significant change in visit duration (25.7 vs. 21.9 min; p = 0.262). Documentation quality improved in accuracy/completeness (p = 0.008), organization/structure (p < 0.001), and overall score (median 23 vs. 17; p = 0.008). Documentation of information provided to the patient was more frequent (80% vs. 20%; p = 0.003). Post-visit review time accounted for 33%-38% of total time. Fifty-five percent of users reported being satisfied or very satisfied, although 67% reported technical issues and 33% reported requiring extensive corrections. Adoption was concentrated in fewer than one-third of trained clinicians. CONCLUSIONS: The AI medical scribe was associated with improvements in interaction and selected domains of documentation quality. However, technical barriers and limited adoption persisted and should be addressed before wider deployment.