Northwell Health’s northern region has nearly completed its ambient AI rollout across acute care settings, with a planned extension to nursing staff by year-end. Albert Villarin, MD, the system’s vice president and chief medical informatics officer, is already looking past documentation. The question driving his attention now is not what ambient AI can do — it is what the next wave of healthcare AI will demand from health systems, clinicians and the patients it serves.
The shift from paper records to electronic health records took American medicine roughly 18 years. Ambient AI documentation reached comparable adoption velocity in a fraction of that time. That compression carries an implication he finds both exciting and sobering: the tools are moving faster than institutional understanding of how to use them.
“Going from no AI to deployment of AI took about eighteen months of demand, but we’re only at the beginning of understanding how we use it,” he said.
The direction Dr. Villarin sees the technology heading centers on the patient rather than the provider. There is an emerging generation of personal AI health assistants — tools designed to monitor an individual’s health data, communicate with the clinical record and serve as a continuous bridge between patients and their care teams. Health systems are already building toward that future, deploying AI agents designed to act as patient companions across scheduling, chronic disease management and post-discharge follow-up.
“We’re creating patient AI bots that will monitor them, monitor their EMR, monitor their life,” Dr. Villarin said. “You’ll have your own personal assistant that does a lot of these factors, and then that will be the bridge or the communication into your EMR for the future.”
The operational payoff he describes extends across the entire care cycle.
“That is not only the burden of the scheduler for an appointment or the patient for the medications or the clinician for following up lab tests,” he said. “It’ll be an automated process.”
Across the industry, the shift from AI tools that assist to agentic tools that act autonomously is accelerating, with health systems expanding deployments into scheduling, call management and clinical workflow support.
Dr. Villarin is neither dismissive of the technology’s pace nor naively optimistic about what it has delivered so far. The enterprise-scale financial return on healthcare AI remains a work in progress across the field. He sees that not as a problem to solve in isolation but as a condition that requires time, stewardship and trust.
“We haven’t seen a large tremendous global enterprise healthcare ROI yet, but we will as we become better stewards of the technology as well as building that trust,” he said. “I’m excited that it’s here, but I’m also excited that we can facilitate trust around it to make the clinical practice of medicine a better place to enjoy.”
Trust is a precondition for the technology to deliver on its promise. He points to the AMA’s joy in medicine framework as a marker of what the AI work is building toward: the idea that returning time and attention to clinical practice is an end in itself, not a byproduct.
The third element of his forward view is the training pipeline. Deploying ambient AI is one challenge; preparing the next generation of clinicians to use it well is another, and the two cannot be separated.
“In the future, I think it’ll be an important part of not only current state clinical practice, but training,” he said. “I’m thrilled to know that we are also training our residents and have a clinical informatics rotation that brings in AI into their education.”
Northwell’s clinical informatics rotation is part of a broader push in academic health systems to integrate AI into residency training — a push grounded in the reality that tools clinicians will use throughout their careers should be part of their foundational education, not discovered in practice.
“Now we have a tool just like a stethoscope needs to be understood, how to practice it, how to use it, and what results come from it, AI is the same way,” Dr. Villarin said. “It’s another facilitated tool that healthcare now can utilize and deliver more advanced care. And it’s our responsibility to train those into the future to bring that better practice through new technologies such as AI.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
