Solving practice challenges with AI

In this episode of the Inside Family Medicine podcast, recorded at FMX, AAFP Vice President of Practice Advancement Karen Johnson talks with family physician Nipa Shah, MD, FAAFP, and internal medicine physician Sudeep Bansal, MD, about how artificial intelligence is moving from an emerging technology to a practical tool physicians can use to solve everyday problems in primary care.

Shah and Bansal discuss how their views of AI have evolved as the technology has rapidly advanced. Shah points to agentic AI – systems capable of performing tasks rather than simply generating information – as an important turning point. Bansal shares how experimenting with relatively small use cases in his own practice changed his perspective from viewing AI as a novelty to seeing it as a tool that can reduce repetitive work and give physicians more time with patients.

The conversation explores some of AI's biggest opportunities and challenges in primary care. Shah describes the potential to use large amounts of health data to improve prevention, diagnosis and treatment, while Bansal focuses on eliminating the many small administrative burdens that make primary care feel like "death by a thousand cuts." They also examine a major obstacle: health information remains fragmented across EHRs, health systems and other data sources. AI may help physicians turn the information they do have into actionable summaries at the point of care, but better interoperability remains essential.

The guests then reflect on the AAFP's first FMX AI Hackathon, where family physicians worked in teams to identify real problems from their practices and rapidly prototype potential solutions. Shah's team developed an AI-supported tool for onboarding, evaluating and providing feedback to medical assistants, while other participants tackled challenges such as helping patients understand what to do after a hospital discharge. Bansal encourages physicians to start with a problem they understand well, use AI to brainstorm possible solutions and remember that clinicians remain the domain experts who must guide the technology.

The episode closes with practical advice for physicians beginning to use AI. Shah suggests expanding the traditional "build versus buy" decision to consider whether to build, buy, borrow or bot a solution. Bansal encourages physicians to experiment with AI on problems that consume their time, while Shah emphasizes transparency about how much AI contributed to a piece of work. Both stress that AI should strengthen rather than undermine the human relationships at the center of family medicine — and that family physicians need a meaningful role in determining how these technologies are developed and implemented.

References and resources

• Artificial Intelligence in Family Medicine
  AI in Family Medicine

• FMX 2026 AI Hackathon highlights

• Getting the right AI tools for your practice: Questions to ask
  https://www.aafp.org/practice-operations/ai-in-family-medicine/choosing-the-best-ai-health-care-tools

Libsyn chapter markers

00:00 AI in primary care and meet the guests
01:35 Sudeep Bansal's path from medicine to health technology
03:39 Nipa Shah on AI as a tool for solving problems
06:59 Why agentic AI is changing primary care
10:09 The biggest opportunities for AI in family medicine
14:28 Data silos, interoperability and actionable information
19:09 Inside the FMX AI Hackathon
23:37 How physicians can turn pain points into AI solutions
25:41 From "build versus buy" to build, buy, borrow or bot
27:37 Using AI responsibly while preserving trust and human connection