Scaling Agentic Automation in Healthcare: Closing the Perceived Risk Gap That Stalls AI Adoption
Insights from Dr. Shervin Majd•December 19, 2025
Healthcare AI initiatives rarely fail because the technology is not ready. They fail because leaders cannot confidently scale what works. The gap between actual risk and perceived risk keeps automation efforts stuck in pilot mode and slows executive decision making. In a HealthAI Collective lightning talk, Dr. Shervin Majd explains how healthcare leaders can close this perception gap and scale agentic automation without compromising safety, reliability, or credibility.
Healthcare organizations rarely reject innovation because the tech doesn’t work. They hesitate because the perceived risk feels bigger than the reward. As Dr. Majd illustrates, this perception gap distorts leadership decisions: what seems like a 20% actual risk is often seen as an 80% gamble.
Leaders, pressed for time, rely on intuition and incomplete information. Meanwhile, innovators and startups often fail to quantify or communicate real risk.
When that gap stays wide, good ideas stall in pilot purgatory and “innovation fatigue” sets in across the enterprise.
Every stakeholder interprets risk through a different lens:
Translating technical or clinical value into these personalized narratives isn’t optional; it’s foundational. When teams don’t do it, uncertainty fills the gap and then it looks like risk.
Majd’s own approach at Providence involved shadowing frontline staff to learn their pain points before building anything. That empathy-driven discovery became a trust multiplier when rolling out products to 51 hospitals and 1,000 clinics.
C-suite buy-in opens doors, but middle management and end users keep the initiative alive.
Dr. Majd emphasized designing both vertical and horizontal engagement strategies:
He cautions startups against focusing only on “who pays the bill.” Instead, treat influence as distributed because lasting adoption depends on every layer’s understanding and confidence.
AI pilots often fail not from poor ideas but from systems that break when adoption doubles.
“Reckless speed kills that trust,” Majd warned.
Healthcare’s complexity demands iteration but not fragility. Majd recommends load-testing for realistic growth scenarios (2×, 5×, even 10×) and investing early in QA and architecture review boards that include both product and scalability experts.
Agility keeps you adaptive but robustness keeps you credible. Neglect either, and you lose momentum the moment success arrives.
Many health systems still view AI initiatives as IT projects and this framing limits their strategic potential.Majd argues that meaningful adoption requires organizational transformation, not just technology deployment.
His playbook involves ruthless prioritization focusing resources on initiatives with committed champions.The result: faster clarity, cleaner execution, and measurable ROI (such as Providence’s $5M+ annual impact from its NoShow program).
When leaders understand both the true risk and the true return, risk-taking becomes rational – not reckless.
Dr. Shervin Majd brings 15+ years of leadership in digital health, AI, and data analytics. As former VP & Chief Innovation Officer at Juvare, he launched Exchange 2.0 with major deployments like Kaiser NorCal. At Providence Health, he scaled AI-driven products across 51 hospitals and 1,000 clinics, including the NoShow program that generated $5M+ annually.He advises startups and institutions such as UCSF, HLTH, and SeamlessMD, and is a member of Life Science Angels. Shervin holds a Ph.D. in Biomedical Engineering.
This article is based on a talk given by Dr. Shervin Majd for HealthAI Collective community.
The Perceived Risk Trap