On July 10, Nabla hosted the inaugural NYC edition of The Ambient AI Rounds at DUMBO House in Brooklyn, welcoming healthcare innovators, clinicians, and technologists. Co-founders Alex LeBrun (CEO) and Delphine Groll (COO) brought together leading voices in AI and healthcare to explore one question: how can artificial intelligence restore the human connection at the heart of medicine? The panel, moderated by Dr. Chethan Sarabu, Director of Clinical Innovation at Cornell Tech, featured Dr. Mark Zhang, Instructor of Medicine at Harvard Medical School and President of AMXRA, and Alex LeBrun.
.jpg)
From paper to presence: ambient AI’s real-world impact
Dr. Mark Zhang opened the discussion with a reflection on his clinical training – a journey from paper records to dictation tools, and now, to ambient AI. The shift, he noted, isn’t just technological. It’s deeply personal. “Seeing patients,” he said, “is not work, but a privilege. The paperwork is the painful part.” For Zhang, ambient AI’s value lies in restoring face-to-face time between clinician and patient – time too often lost to documentation.
Alex LeBrun echoed this, sharing a personal story from the field: when a clinic's IT system went down and clinicians were, paradoxically, delighted to return to pen and paper. “For the last 50 years,” he observed, “IT has largely added burden to clinicians. EHRs were built for billing and compliance, not for care. With AI, we have a chance to go back to the joy of care.”
From buzzword to building block: what ‘agentic AI’ really means
While “agentic AI” may be the latest industry buzzword, the panelists stressed that its implications are both practical and systemic. If ambient AI focused on passively capturing and structuring information, agentic AI is about intelligent action: adapting to real workflows, making recommendations, and responding to context.
“In the past, we built one-size-fits-all tools,” said LeBrun. “Agentic AI allows for tailored solutions. Agents are like prompt-driven mini-programs that clinicians can compose and combine to reflect real-world, nuanced workflows.”
Dr. Zhang described this as a step toward AI systems that support clinicians more like teammates than tools. “We’ve gone from dictation to ambient scribing, and now we’re approaching agents that function almost like junior residents: assisting, routing, recommending,” he said. “But we’re just at the beginning.” For now, he emphasized, clinician oversight remains essential to ensure both safety and effectiveness.
AI’s true north: patients
The conversation also addressed a critical dimension often overlooked in technical discussions: the patient experience. LeBrun shared that physicians using ambient AI see, on average, a 10% increase in patient satisfaction, a metric he attributed to something simple but powerful: presence. When clinicians are freed from screens and documentation, they can focus fully on the patient in front of them. That same principle extends beyond the visit. “Eighty-five percent of what a doctor says is forgotten,” LeBrun noted. “With AI-generated visit summaries, patients retain more, understand better, and feel seen.”
These summaries, automatically generated and easily shareable, reinforce communication between visits, especially for patients who may feel overwhelmed or anxious during appointments.
Zhang added that AI could also help realize the long-standing but elusive goal of a true personal health record, one that is interoperable, patient-owned, and dynamic. “All the value AI has unlocked for clinicians,” he said, “can be doubled or tripled for patients.”
A cautious path forward
Still, the panel was clear-eyed about the challenges ahead. Zhang acknowledged that while AI brings new capabilities, avoiding the pitfalls of past technologies, like EHRs, will require more than better tools. It will take rethinking how care is delivered, and who it’s designed for. “We're good at iterations,” he said. “But systemic change may depend as much on societal expectations as on healthcare itself.”
LeBrun noted that early excitement around AI has been driven by “soft ROI”: improvements in clinician well-being, time saved, and reduced burnout. But as the technology matures, expectations will rise. Systems will need to deliver measurable impact: improved coding accuracy, reduced documentation time, and, ultimately, better health outcomes.
Rebuilding, not just optimizing
Both panelists agreed that while AI is helping optimize current workflows, meaningful transformation will require going beyond patchwork fixes. “We’re only 1% into this journey,” LeBrun said. “If I had $10 billion, I wouldn’t just add AI to the system, I’d rebuild it from the ground up.” For now, he added, the task is to be thoughtful in where AI is applied and bold in imagining where it could lead.
Zhang echoed the need for intention: “AI will be part of everything we do, whether we like it or not. But we have to move forward consciously.”
A collective mission
The discussion underscored a central takeaway: AI is not just a technology, it’s an inflection point. Its success depends on thoughtful, interdisciplinary collaboration among clinicians, technologists, and patients. As one audience member asked: “How will AI help us advance humanity in medicine?” The answer, the panel suggested, lies not in the models themselves, but in how we choose to use them – to restore time, connection, and empathy to the center of care.
Watch the recording and stay tuned for more from Nabla as we continue advancing our mission to restore the human connection at the heart of healthcare through industry-leading clinical AI.
