Deutschland
India
Lietuva
United Kingdom
Search
Search
Search
Healthcare leaders have moved past speculation about artificial intelligence and into a more practical phase: identifying where AI can improve care, reduce administrative burden, and scale safely across complex organizations.
That shift was a central theme of a panel at Guidehouse’s 2026 Healthcare Innovation Summit moderated by Guidehouse Director Alex Holston, MD. Speakers included:
Holston, a practicing neonatologist, opened the discussion by grounding AI’s promise in the lived reality of clinicians and patients. He described teaching an AI agent his clinical workflows and using it to help him draft notes, giving him back as much as 20 to 30 minutes of additional time per patient.
“What does that 20 to 30 minutes mean? It doesn’t mean I get to sit in the call room and watch more Netflix,” Holston said. “What that means is I’m available to talk to my patients’ families.”
For Holston, this example illustrates the broader stakes. AI-enabled efficiency is not only about productivity, he said, but about giving clinicians more capacity to reassure families, explain complex care decisions, and focus on the people behind every clinical interaction.
Panelists pointed to several areas where AI is already delivering measurable improvements. At the Veterans Health Administration (VHA), Carey said more than 200 AI products are already deployed across the organization.
One example is a chatbot designed to help users navigate the health system’s newly implemented EHR. Training employees across VHA’s 1,400 facilities is a massive, multi-step effort, so staff created the chatbot as a bridge solution to help staff get acclimated with the system faster. The chatbot has access to training materials about the EHR and can quickly source answers for staff about how to use the new system.
“You can just go to this chatbot and ask essentially any question about the new EHR,” Carey said. “In the absence of a chatbot like this, staff would be saying, ‘It's on page 36 of that PDF you got in the fourth hour of that all-day training session you went to,’ which of course isn’t very helpful.”
AI is also helping VA care coordinators expedite the extremely manual effort of processing documents for care received by Veterans outside of VA, and supporting primary care clinicians with ambient scribes that allow them to focus on patient conversations.
Shaikh highlighted how Amazon Connect Health is using agentic AI to support patient identification, insurance verification, and scheduling across millions of patient contact center calls across the country. The company also worked with a large Northeast health system to improve outcomes for high-risk diabetes patients by using AI to synthesize clinical, social, and device data—helping clinicians identify preventable risks and intervene earlier.
Radiology Partners is using AI across its business through its proprietary Mosaic Clinical Technologies platform to strengthen support services like its contact center, revenue cycle management, and other key areas. These tools are helping the organization measurably improve productivity for providers and staff working across the organization.
“We’re seeing nice productivity improvements through the use of our platform for radiologists,” Graves said. “And if you think about what that means for backlogs and capacity across the industry, it is meaningful and will be a gamechanger.’
Moskowitz, who leads Bain’s Healthcare IT and MedTech investing, described AI’s potential across two major fronts: expanding clinical capacity and reducing administrative costs. He said long waits for specialty care and rising administrative spending have created an urgent need for new approaches.
“As we all know, healthcare is the largest part of the U.S. GDP,” Moskowitz said. “Twenty percent of the $5 trillion we spend is spent on administrative costs.”
That administrative complexity makes healthcare especially ripe for AI-enabled automation, he noted, in areas such as revenue cycle management, contact centers, scheduling, and patient navigation.
Even with these successful use cases, more governance is needed to realize AI’s full potential, panelists said.
Implementing AI tools requires a different approach than traditional IT implementations, Shaikh noted, particularly when they begin to reshape clinical and operational workflows. That means organizations need role-specific training that helps employees understand not only how to use AI but how their processes may need to change. Unlike traditional automation, which often helps people complete existing tasks more efficiently, agentic AI can require organizations to rethink how work gets done.
“There’s a general level of AI literacy that needs to be instilled in an organization,” Shaikh said. “You need to have a process that allows you to, from the C-suite to entry-level employees, look at the development and implementation—how it changes processes and adds value to each role.”
Governance at VA includes a publicly available AI use case inventory, ongoing risk assessments, and mandatory informed consent for tools that directly impact patient care.
“Our most important asset, I would argue, is Veteran trust,” Carey said. “We have to steward that trust as we look to innovate in this AI space.
Panelists also cautioned that not every AI idea is worth pursuing. Moskowitz said organizations must be willing to experiment quickly, tolerate some failures, and focus investment on the highest-value opportunities.
“Move fast and don’t be afraid to fail,” he said. “You’re really going to have to embrace that not every solution is going to work.”
He added that the strongest opportunities are those tied to large labor pools, unresolved customer problems, or workflows where automation can clearly reduce costs, expand capacity, or improve productivity.
Smaller productivity gains may not be enough to justify AI investment if they don’t translate into meaningful capacity, cost, or workflow improvements. “We’re pointing our AI projects at our biggest business problems,” Graves said.
To test solutions, Graves said his team has run time-bound comparisons among technology partners and internal teams to determine who can build the most effective product fastest. In some cases, that has meant choosing a partner when speed to deployment outweighed the preference to build internally.
That discipline—moving quickly while staying focused on measurable outcomes—is what panelists agreed will set forward-thinking organizations apart. AI’s future in healthcare will depend not on experimentation alone but on whether organizations can identify the right problems to solve, then align technology, workflows, and trust around them.
Guidehouse is a global AI-led professional services firm delivering advisory, technology, and managed services to the commercial and government sectors. With an integrated business technology approach, Guidehouse drives efficiency and resilience in the healthcare, financial services, energy, infrastructure, and national security markets.