Here are five takeaways from the event, according to Becker’s interviews with executives on site:
1. The AI race. Many health system leaders discussed their enterprise rollouts or pilots of ambient AI technology to document patient visits. “I’ve been in this business 30 years and nothing has come close to being this impactful, this quickly, at the right time,” said Eric Poon, MD, chief health information officer of Durham, N.C.-based Duke University Health System, which turned on Abridge for 5,000 providers and trainees in January.
But he noted that a lot of AI solutions look promising but don’t end up working, for reasons both technological and human. “With AI, you need to kiss a lot of frogs to find your prince or princess,” he said.
At a panel, Veena Jones, MD, vice president and chief medical information officer of Sacramento, Calif.-based Sutter Health, called implementing ambient AI “one of the highlights of my career so far.” Next week, Somerville, Mass.-based Mass General Brigham plans to offer it to 7,000-plus providers. “It’s not cheap, but it’s transformative,” said Rebecca Mishuris, MD, chief medical information officer and vice president of digital at Mass General Brigham, during the panel. “Burnout really is a crisis. It is costly to replace a physician or APP [advanced practice provider] or … when they pull back on their time.”
“The ultimate goal of healthcare AI is to make me a better doctor,” said Brian Hoberman, MD, executive vice president of IT and CIO at the Permanente Federation, part of Oakland, Calif.-based Kaiser Permanente. Ambient AI, he said, has done that for him and his colleagues.
Many health system leaders are excited about the technology moving on to nurses (with some trialing it for nursing documentation).
2. Agentic AI. AI agents have arrived in healthcare. An AI agent at Charleston, S.C.-based MUSC Health is completing 40% of prior authorizations without human intervention. Epic is among the companies developing AI agents to, say, identify and close care gaps.
3. Doing care at home differently. Mass General Brigham is expanding hospital at home to post-operative and oncology patients. The health system is also developing a care orchestration platform with Best Buy Health to streamline the at-times overwhelming logistics needed to run the care model. Stephen Dorner, MD, chief clinical and innovation officer of Mass General Brigham Healthcare at Home, said he envisions a role for AI in not just detecting falls, but preventing them, by learning about each specific patients’ movements and habits at home.
Palo Alto, Calif.-based Stanford Health Care wants to do inpatient treatment in the home, but California regulations don’t allow it. So the health system plans to launch advanced care at home May 1, using AI to identify patients and wearables and remote patient monitoring technologies to treat them. Stanford aims to become a statewide brand for home care as its academic medical center serves patients from Fresno to the Oregon border. “We’re not trying to take a population health approach,” said Gretchen Brown, MSN, RN, vice president and chief nursing information officer of Stanford Health Care. “We’re trying to send pretty complex patients home.”
4. Beyond ROI. Emerging technologies like AI might not have immediate financial returns. “We need to be OK with not worrying about ROI. We’re not going to have ROI right away,” said Crystal Broj, chief digital transformation officer of MUSC Health. “The AI’s learning. We’re learning.”
For instance, when MUSC Health started testing Microsoft’s DAX Copilot AI scribe a year ago, it didn’t work all that well. By the end of the pilot, providers didn’t want MUSC Health to take it away. “You have to go through the bad to get to the good,” Ms. Broj said.
5. Don’t forget about cybersecurity. “The focus has shifted to AI, but let’s not forget about cybersecurity,” said Zafar Chaudry, MD, senior vice president and chief digital, AI and information officer of Seattle Children’s. “I truly don’t believe that in 2025 we’re any safer than we were in 2024.”
AI can both help improve cybersecurity, by detecting threats, and make health systems more vulnerable, like if an employee accidentally uploads malicious code, Dr. Chaudry said.