Many medical practice leaders opted to not be early adopters of AI tools, preferring to wait for technology to mature and be proven effective in other organizations. However, 2025 marks a shift, with AI tools now the top technology priority for medical group leaders.

So how much AI is currently involved in patient care? An Aug. 5, 2025, MGMA Stat poll found that 71% of practice leaders report some use of AI for patient visits, while 29% say AI plays no role in their visits. Importantly, not all practices using AI apply it uniformly across all encounters:
- Nearly half (47%) use AI in 25% or less of all visits.
- About 10% report AI use in 26% to 50% of visits.
- About 7% report AI involvement in 51% to 75% of visits, and another 7% say it’s used in 76% or more of visits.
The poll had 244 applicable responses.
Among practices using AI in patient visits, opinions were divided on its impact on staff workload: About 44% said it has not reduced workload, 39% said it has, and 17% were unsure. Practices reporting broader AI use — in 26% or more of all patient visits — were more likely to indicate a positive impact on workload.
Practices with the most limited AI use noted early improvements in scheduling, appointment reminders, documentation efficiency, and after-hours patient communication. Conversely, those not seeing workload reductions often cited added complexity or being in early stages of AI adoption.
Respondents who experienced workload reductions commonly credited real-time note creation during patient visits, which reduced the need for manual typing and documentation.
Among practices not currently using AI, views on future adoption were mixed. Several leaders expressed clear intent to adopt AI tools in the next year, particularly for scribing, scheduling, phone call management, and EHR-integrated solutions. Others remained cautious, citing barriers such as high costs, limited EHR integration capabilities, unclear return on investment (ROI), accuracy concerns, or lack of organizational readiness.
Where AI is being deployed today
One of the most common use cases is documentation and coding automation through ambient listening technology, which can reduce clinician documentation time. Additionally, AI tools are increasingly being used in revenue cycle management — particularly for prior authorization management, eligibility checks, and denial prediction — to streamline administrative workflows and improve financial performance.
Patient access and triage tools, such as chatbots, symptom checkers, and automated scheduling systems, are also gaining popularity. These tools are especially helpful for practices seeking to enhance patient engagement during evenings or weekends.
- For more, read MGMA’s 2025 Artificial Intelligence Issue Brief (PDF).
Roadblocks and mitigation
Cost and unclear ROI are commonly cited barriers to AI adoption beyond this week’s polling. Many practices can mitigate these concerns by launching short-term pilot programs focused on a specific goal — such as reducing time spent on documentation — and closely tracking the results before scaling up.
A joint report by Humana and MGMA in late 2024 identified additional challenges: insufficient education/training, lack of trust/buy-in, and limited internal capacity to support implementation. Concerns about compliance and data privacy were also noted, though these issues tended to diminish once AI was in place and understood.
Developing a lightweight AI governance structure and selecting vendors based on clear interoperability, regulatory compliance, and security criteria can help address many of these concerns.
- Read more about the recent AI Action Plan released by the White House.
Sample timelines for AI deployment in ambulatory care
First 0-6 months:
- Form an AI governance group to oversee vendor selection, compliance, and risk management.
- Standardize AI vendor evaluations using criteria such as FHIR-based interoperability, HIPAA and SOC 2 compliance, FDA regulatory status, and security standards.
- Potential use cases:
- Pilot ambient voice-scribe solutions in select exam rooms to cut documentation time.
- Integrate patient-facing AI tools, such as chatbots, into portals for scheduling or triage.
- Automate revenue-cycle processes, especially prior authorizations or basic coding.
6–12 months:
- Expand ambient scribe use across additional clinicians and standardize note templates.
- Implement staff training led by AI "super-users" to accelerate adoption and instill confidence.
- Introduce predictive analytics for scheduling improvements (e.g., no-show risk scoring, slot optimization).
Metrics to track your organization’s progress:
- Ambient scribes: Shorter time to complete visit notes.
- Patient-facing AI: Faster response times in scheduling and portal messaging; higher same-day appointment fill rates.
- AI for RCM: Less time spent on prior auths and appeals; fewer denied claims.
Specialty-specific AI considerations in Year 1:
- Primary care specialties: Deploy chronic condition dashboards using predictive analytics for diabetes, hypertension, or other common conditions. Track improvements in clinical markers such as A1c levels and blood pressure control.
- Surgical specialties/ASCs: Evaluate and adopt diagnostic imaging tools (e.g., AI-assisted polyp detection or ultrasound guidance). Consider OR scheduling and/or supply-chain forecasting tools to optimize throughput and reduce waste.
- Nonsurgical specialties: Implement AI diagnostic tools (e.g., ECG, dermatology, retinal scan interpretation). Use predictive risk models to support population health strategies (e.g., heart-failure readmission in cardiology).
Join MGMA Stat
Our ability at MGMA to provide great resources, education and advocacy depends on a strong feedback loop with healthcare leaders. To be part of this effort, sign up for MGMA Stat and make your voice heard in our weekly polls. Sign up by texting “STAT” to 33550 or visit mgma.com/mgma-stat. Polls will be sent to your phone via text message.