A special report on AI in primary care from Annals of Family Medicine

Author(s):

Austin Littrell

Key Takeaways

AI can help reduce primary care physician burnout by addressing time-consuming tasks, especially in EHR management.
Effective AI tools must focus on specific challenges, such as documentation and medication management, to deliver tangible time savings.
AI alone cannot solve systemic issues like excessive patient panels; structural changes are also necessary.
Successful AI implementation requires healthcare organizations to prioritize clinician well-being and patient care.

A new report in Annals of Family Medicine urges AI developers to target time-saving AI solutions—and warns against overpromising.

As the primary care industry grapples with high burnout rates and unsustainable workloads, many physicians have turned to artificial intelligence (AI) as a potential lifeline. A special report in the January 2025 issue of Annals of Family Medicine contends that, in order for AI to successfully transform primary care, the industry must focus on solving specific, pressing issues—beginning with clinicians’ time.

The report, authored by John Thomas Menchaca, MD, of the University of Utah, argues that AI tools will fail unless designed to tackle the right challenges. “The key will be to start with the right problem before jumping to solutions,” Menchaca writes, likening poorly targeted innovations to the Segway, a notoriously overhyped transportation tool that never truly met consumer needs.

Instead, Menchaca points to the success of rentable scooters—narrowly tailored to address the “last mile” challenge of urban commutes—as a model for AI’s potential. The primary care equivalent of this “last mile” issue, Menchaca suggests, is time: clinicians spend too much time working, particularly dealing with electronic health records (EHRs).

Time and physician burnout

Menchaca points to a 2017 Annals of Family Medicine study, which determined that, on average, full-time primary care physicians spend more than 11 hours daily on work, with more than half of that time devoted to EHR tasks, including documentation, chart reviews and patient messaging. This workload, specifically extended time spent in EHRs, has been repeatedly linked to higher rates of burnout.

Menchaca identifies documentation as the most promising area for AI intervention. AI-powered tools, like automated transcription and note generation, are already emerging, with the potential to cut hours off clinicians’ daily routines. However, the report cautions that early iterations of these AI tools often fall short. Clinicians at the recent Artificial Intelligence in Medicine (AIME) conference expressed frustration that AI-generated notes often require substantial editing, undermining their time-saving intentions.

Additional targets for AI

Beyond documentation, Menchaca argues that other EHR-related tasks could benefit from AI solutions. Medication management, for instance, stands out as a key opportunity for AI innovation. Menchaca highlights research demonstrating that clinicians with in-clinic pharmacy technicians spend eight fewer minutes in the EHR per visit, working on tasks including medication refills, fill issues and prior authorizations. Scaled across a typical workday, these savings could amount to two hours.

Avoiding the “segway effect”

The report stresses that AI tools should focus on the tasks that consume the most time for primary care physicians, particularly within EHR. Menchaca notes that past tools, including clinical decision support systems, failed because they increased rather than reduced clinician workload. For AI to succeed, it must deliver tangible time savings—anything less risks making it another overhyped, yet underperforming, innovation.

Menchaca also reminds clinicians that AI alone cannot solve systemic problems in primary care, including excessive patient panels and overburdened schedules. Although AI is capable of streamlining workflows, it cannot address the organizational issues that drive burnout. Meaningful improvements will require both technological solutions and structural changes to prioritize clinician well-being.

“AI will not shrink ballooning patient panels or outmaneuver overloaded schedules,” Menchaca wrote. “Technology can only be as effective as the system in which it operates, and primary care clinicians will only reap the benefits of AI if it is implemented in organizations that sincerely prioritize clinician well-being and patient care.”

Ultimately, Menchaca argues that AI is not an end in itself, but a tool to lighten the administrative burden in primary care. Whether it fulfills this promise is dependent on how well it is implemented to meet the needs of clinicians.