Five real-world examples. Four questions to guide your next technology decision. Explore the roundup
Written by: American Hospital Association • Created: Sep 17, 2026
Daniel Smith, PA-C, EMBA Chief Medical Information Officer Corewell Health
Clinical AI, informatics governance and EHR innovation
Responsible AI adoption that reduces clinician burden and creates more time for patient care
Daniel Smith is Chief Medical Information Officer at Corewell Health, where he leads enterprise clinical informatics and guides the responsible adoption of AI, digital health technologies and electronic health record innovation across one of the nation’s largest health systems.
A physician assistant by training, Smith moved into informatics to improve care beyond the individual patient encounter. As he became more involved in EHR design and optimization, he saw a persistent problem: clinicians were spending more time serving the technology than being supported by it.
That experience now shapes his approach to AI. For Smith, success is not about adding more technology, but about using it to reduce friction, strengthen clinical decision-making and give clinicians more time to focus on patients.
I started as a physician assistant in emergency and primary care and wanted to have impact beyond individual patients.
That led me into EHR build work, then clinical informatics, and eventually a broader leadership role. What began as a one-year opportunity became a career.
What kept me here was seeing how much clinicians were struggling with technology. I originally wanted to help patients navigate their care better. Over time, that expanded to helping clinicians do the same.
Today, I focus on making sure the tools we build actually support care delivery instead of getting in the way.
The biggest opportunity is making technology less visible during care, not more.
Clinicians are still tied to screens for documentation and administrative work. I want a future where they can focus fully on patients while technology works in the background.
We are starting to see that with ambient tools, which help reclaim time and reduce distraction.
But automation alone is not enough. If we only digitize existing processes, we risk reinforcing inefficiencies.
The real value is in augmentation, surfacing the right information at the right time and supporting better clinical decisions so clinicians can focus on judgment and care.
We try to move quickly, but carefully.
AI is still evolving, so we focus on low-risk, high-value use cases like reducing administrative burden.
We also protect the first experience. If a tool performs poorly early on, it is hard to rebuild trust.
We have stopped pilots when technology was not ready, even if we believed it would be valuable later.
At the same time, we continuously monitor what we deploy. For example, we refined an inpatient summarization tool after identifying issues in how information was being interpreted.
These systems require ongoing testing, feedback and adjustment. You cannot implement and walk away.
It starts with frontline buy-in.
We do it with clinicians, not to them. The goal is not to decide what is best for physicians and then ask them to adopt it. We involve them early in defining the problem and designing the solution so it reflects what they actually need.
That matters because not every problem needs AI or a brand new tool. Even a promising solution will struggle if it does not fit the way clinicians work.
When clinicians see real value, they become the strongest advocates. That peer influence is what drives broader adoption.
Health systems are often solving similar problems in isolation.
I regularly connect with peers to learn what is working and what is not, sometimes even visiting sites in person. That learning can prevent costly missteps.
The Accelerator Champion Network helps accelerate that exchange. I joined to learn from others and avoid repeating avoidable mistakes, while also sharing our own experience.
In a fast-moving space like AI, trusted peer insight helps separate real value from hype.
Ultimately, the goal is shared progress so patients everywhere benefit from better care, not just those in early-adopting systems.
Health systems do not have to absorb all the financial and operational risk of evaluating emerging technology.
Smith believes organizations should expect vendors to demonstrate value before major investment in integration and implementation.
Traditional pilots often require significant time from technology teams, workflow redesign, contracts and clinical staff before it is clear whether a product works, and that early investment can create pressure to continue even when results are weak.
Where possible, hospitals should start with lower-risk evaluations, asking vendors to demonstrate capabilities in low-integration environments and defining success criteria up front.
It also is important to account for the full cost of adoption, including implementation, workflow change, training, maintenance, model updates and ongoing monitoring.
Ultimately, technology should be judged by what it returns: more time for clinicians, greater capacity to care, improved outcomes and more space for human connection.
Explore how Corewell Health is using ambient AI to reduce documentation burden and give clinicians more time to focus on patients.
Explore the solution
Meet additional leaders sharing practical experience to help hospitals evaluate, implement and scale technology with greater confidence.
Meet the Champions