Utilizing Artificial Intelligence (AI) in the Primary Care Setting: A Quality Improvement Project

Abstract

Purpose: Primary care provider (PCP) burnout is largely driven by electronic health record (EHR) documentation burden. This project implemented a toolkit to support the sustainable use of an ambient artificial intelligence (AI) documentation tool (Heidi) to improve efficiency and provider experience in primary care. Background: PCPs spend nearly half of their clinical workday on EHR documentation, contributing to reduced patient interaction, inefficiency, and burnout. National organizations have identified clinician burnout as a public health crisis and emphasize system-level interventions. Ambient AI documentation tools offer a promising solution; however, structured implementation strategies in primary care remain limited. Sample/Setting: This quality improvement project was conducted in an urban, midsize family medicine clinic in the Pacific Northwest. Participants included six PCPs, nurse practitioners, physician assistants, and physicians, using or preparing to use Heidi within the electronic health record system. Methods: A comprehensive toolkit was developed, including provider education, standardized documentation templates, AI-related policy guidance, and evaluation tools. The project was implemented over 12 weeks using the Plan-Do-Study-Act (PDSA) framework to guide iterative intervention and evaluation. Results: Providers reported improved satisfaction with documentation workflows and perceived reductions in documentation burden following toolkit implementation. Standardized templates improved documentation consistency and efficacy. Qualitative feedback indicated increased focus on patient care, improved workflow efficacy, and greater confidence in the ethical use of AI documentation tools. The toolkit was feasible and acceptable for integration into primary care. Conclusion: A structured implementation toolkit supported the successful adoption of ambient AI documentation in primary care. Education, standardized workflows, and policy guidance contributed to improved provider experience and perceived efficiency. Further research is needed to evaluate long-term effects on burnout, efficiency, and patient care outcomes.

Keywords: primary care, burnout, electronic health records, ambient artificial intelligence, Heidi, documentation burden, quality improvement