For AI in Primary Care, Start With the Problem

John Thomas Menchaca

doi:10.1370/afm.240504

Article Figures & Data

Additional Files

PLAIN-LANGUAGE SUMMARY

Special Report

Before Implementing AI in Primary Care, Understand Where Clinicians Need the Most Help

Background and Goal:Primary care clinicians face significant burnout, driven by excessive administrative tasks and time spent on electronic health records (EHRs). This report emphasizes that generative AI tools must focus on addressing specific, impactful problems. Drawing on the failure of the Segway, the report argues that AI must avoid becoming a "solution looking for a problem."

Key Insights:The Segway, once expected to revolutionize transportation, failed because it did not solve a real need. Rentable scooters succeeded by addressing a narrow, specific problem: the “last-mile” challenge in urban commutes. Similarly, AI in primary care must tackle clinicians' “last-mile” issue—time. With over half of their 11-hour workdays spent on EHR tasks, clinicians need AI to target key areas like documentation, chart reviews, medication management, and patient communications. Collaboration between clinicians, innovators, and researchers is critical to ensure AI tools address real-world needs. Systemic challenges—such as overloaded schedules and ballooning patient panels—cannot be solved by technology alone. For AI to succeed, health care organizations must prioritize clinician well-being and align tools with practical needs.

Why It Matters:The failure of the Segway illustrates the risk of innovating without understanding the underlying fundamental problems that need to be addressed. AI has the potential to reduce primary care burdens and improve work-life balance, but only if implemented thoughtfully. Technology works only as well as the system in which it operates, and primary care clinicians will only reap the benefits of AI if it's implemented in organizations that prioritize clinician well-being and patient care.

For AI in Primary Care, Start With the Problem

John Thomas Menchaca, MD

Internal Medicine and Biomedical Informatics, University of Utah, Salt Lake City, Utah
PLAIN-LANGUAGE SUMMARY

Special Report

Before Implementing AI in Primary Care, Understand Where Clinicians Need the Most Help

Background and Goal:Primary care clinicians face significant burnout, driven by excessive administrative tasks and time spent on electronic health records (EHRs). This report emphasizes that generative AI tools must focus on addressing specific, impactful problems. Drawing on the failure of the Segway, the report argues that AI must avoid becoming a "solution looking for a problem."

Key Insights:The Segway, once expected to revolutionize transportation, failed because it did not solve a real need. Rentable scooters succeeded by addressing a narrow, specific problem: the “last-mile” challenge in urban commutes. Similarly, AI in primary care must tackle clinicians' “last-mile” issue—time. With over half of their 11-hour workdays spent on EHR tasks, clinicians need AI to target key areas like documentation, chart reviews, medication management, and patient communications. Collaboration between clinicians, innovators, and researchers is critical to ensure AI tools address real-world needs. Systemic challenges—such as overloaded schedules and ballooning patient panels—cannot be solved by technology alone. For AI to succeed, health care organizations must prioritize clinician well-being and align tools with practical needs.

Why It Matters:The failure of the Segway illustrates the risk of innovating without understanding the underlying fundamental problems that need to be addressed. AI has the potential to reduce primary care burdens and improve work-life balance, but only if implemented thoughtfully. Technology works only as well as the system in which it operates, and primary care clinicians will only reap the benefits of AI if implemented in organizations that prioritize clinician well-being and patient care.

For AI in Primary Care, Start With the Problem

John Thomas Menchaca, MD

Internal Medicine and Biomedical Informatics, University of Utah, Salt Lake City, Utah