Clinicians Say Ambient AI Is Cutting Documentation Time and Reclaiming Patient Care Time

# What clinicians are reporting Clinicians at several health systems describe ambient AI—tools that listen to encounters and help generate clinical documentation—as reducing time in electronic health records and restoring time for direct patient interaction. Health systems featured include Penn Medicine, Temple Health and Advocate Health. They report measurable time savings, new workflow questions and the need for careful implementation.

# How systems are measuring outcomes Penn Medicine has tracked quantitative results and qualitative feedback. Early data show a roughly 20% reduction in documentation time and decreased after-hours documentation for 79% of users. Those metrics also raised other signals: an increase in orders placed after clinicians began using ambient tools, which Penn Medicine is investigating as a possible sign of more thorough patient conversations.

Temple Health's chief medical information officer, Dr. Benjamin Slovis, frames ambient AI as promising but not a cure-all. He says the organization has "a lot of data to suggest that we are heading in the right direction," underlining the need to collect outcomes beyond simple adoption rates.

# Tailoring technology for roles and settings Advocate Health is rolling out Microsoft Dragon Copilot to physicians and is now adapting it for nurses. The organization's leaders emphasize that physician and nurse documentation patterns differ, so a single configuration won't fit every role. Penn Medicine also customized ambient AI use across ambulatory, inpatient, emergency and home care, and is testing specialty clinician platforms and charting for home health.

# Implementation and change management All three organizations stress multidisciplinary implementation teams and staged testing. Penn Medicine highlights the need for collaboration among clinicians, operations, data science, ethics, privacy, security, infrastructure and informatics to make deployments sustainable. Training, gradual implementation and explicit guidance for clinicians on how to use the technology and explain it to patients were recommended.

Temple Health and Penn both point to ongoing improvement cycles: collect data, identify unexpected effects (like changes in order volumes), adjust workflows and expand carefully. That approach seeks to preserve clinical judgment and patient trust while reducing clinician administrative burden.

# Early qualitative signals

# Practical implications for later adopters

• Measure a range of outcomes: documentation time, after-hours work, downstream clinical actions (orders), and patient experience.
• Customize configurations for clinician roles and specialties rather than applying a single setup across the enterprise.
• Use staged rollouts with training and feedback loops to identify friction and iterate on workflow design.
• Assemble cross-functional teams that include clinical, technical, privacy and operational experts to oversee deployment and ongoing monitoring.

# Bottom line Early adopters report tangible reductions in documentation time and improved clinician experience, while also uncovering operational questions that require data-driven follow-up. Success depends on role-specific tailoring and continuous improvement rather than a one-time technology drop-in.