Augmented Intelligence at the Point of Care: From Hype to Daily Practice

News | Published: Tuesday, March 31, 2026

---

At the Louis F. Bishop Keynote during the ACC.26 “Artificial Intelligence (AI) Intensive I: Augmented Intelligence at the Point of Care – From Hype to Daily Practice,” Mintu Turakhia, MD, MS, chief medical and scientific officer + EVP Advanced Technologies, iRhythm, and professor of Medicine and Cardiac Electrophysiologist at Stanford University, emphasized a pivotal shift in the use of AI in cardiovascular care. The challenge is no longer innovation but implementation. While the past decade has seen rapid advances across multiple waves of AI development, only a small number of models have made it into reimbursed, real-world use. “The bottleneck is no longer discovery. It’s transformation and execution,” Turakhia noted.

A central theme was the lack of infrastructure needed to support AI at scale. The Apple Heart Study demonstrated how quickly a nationwide digital clinical platform could be built—but also how easily it can disappear once a study ends. Despite major technological progress since then, healthcare systems still lack a consistent, unified foundation to deploy AI tools effectively. Turakhia argued that AI must be treated as core infrastructure, much like electronic health records, with strong integration into clinical workflows and fewer, not more, clicks for clinicians.

The keynote also highlighted AI’s growing capabilities in areas like electrocardiogram analysis, where algorithms can now match or even outperform cardiologists in certain tasks. However, these tools often replicate predictable human errors and are approaching a performance ceiling, reinforcing the need for careful oversight. More importantly, high-performance metrics alone are not enough. AI must produce actionable insights that can guide care, such as identifying patients at risk for complications or enabling earlier detection of disease.

As AI expands into workflow automation and clinical decision support, questions around safety, governance and incentives are becoming more urgent. Speakers pointed to the rise of AI-driven tools for scheduling, follow-up and even autonomous clinical tasks, while cautioning that adoption must be paired with strong oversight and clear value. Generating the right kind of evidence – focused on real-world implementation rather than just clinical trials – will be critical.

Ultimately, the session underscored that AI’s future in cardiology depends on execution. With sensing and reasoning capabilities already in place, the next phase will require building the systems, incentives and trust needed to bring AI fully into everyday patient care.

« Back