By Dr. Ed Lee
Ed Lee, MD, MPH, is a practicing internal medicine physician, physician executive, and nationally recognized leader in clinical informatics and AI.
As Chief Medical Officer at Nabla, he leads clinical strategy and innovation for the company's AI-powered documentation platform, helping health systems improve clinician experience and care delivery.
Previously, Dr. Lee served as Chief Information Officer and Associate Executive Director at Kaiser Permanente, where he led enterprise digital transformation initiatives spanning AI governance, telehealth, and clinician-facing technologies. He also served as Chair of Clinical Education and Director of Clinical Informatics at California Northstate University College of Medicine, overseeing technology-enabled medical education and clinical informatics.
A few years ago, the biggest question facing health systems was whether AI belonged in healthcare. Today, that question has changed.
Most organizations have moved beyond experimentation. They're evaluating—or using—ambient documentation, revenue cycle automation, coding, scheduling, clinical decision support, nursing tools, and countless other AI applications. The challenge is no longer finding AI, it's making those investments work together in a way that delivers long-term value.
That's why leading organizations are shifting from selecting individual AI tools to building enterprise AI strategies.
The Problem with Tool-First Thinking
Health systems never set out to create fragmented AI environments. Instead, it happens incrementally. One department purchases an ambient documentation solution, another invests in coding automation, and a third adopts AI for scheduling or patient communications. Each decision makes sense in isolation, but together they can create disconnected clinician experiences, duplicate governance processes, multiple vendor relationships, and inconsistent approaches to implementation and measurement.
The result isn't necessarily better AI. It's simply more AI.
An enterprise AI strategy takes a different approach. Rather than evaluating technologies one by one, it starts with a broader vision for how AI will support the organization's clinical, operational, and financial goals over time.
Start with Organizational Priorities
Technology decisions should be driven by strategic priorities, not the other way around. Before evaluating individual AI solutions, health systems should align on the problems they’re trying to solve and the outcomes they're trying to achieve. Common enterprise goals include:
- Reducing clinician burnout by minimizing administrative burden
- Improving documentation quality to support better clinical decision-making
- Strengthening revenue integrity through more complete, accurate documentation
- Expanding patient access by increasing clinician efficiency and capacity
- Enhancing the patient experience by giving clinicians more time to focus on care
These priorities should become the lens through which every AI investment is evaluated. When organizations begin with shared objectives, it's much easier to determine which technologies fit into a larger roadmap and which simply solve isolated problems.
Design Around Workflows, Not Features
Clinicians don't experience AI as a collection of products. They experience it as part of their daily workflow.
Take clinical documentation. A behavioral health clinician may need AI that captures nuanced, longitudinal patient conversations and produces specialty-specific note structures. An emergency physician, on the other hand, is focused on speed, rapid documentation, and efficient handoffs. While their workflows are very different, both clinicians are trying to accomplish the same goal: accurate documentation that supports patient care without adding administrative burden.
The goal isn't to standardize how clinicians practice. It's to create a unified AI strategy that supports specialty-specific workflows while maintaining a consistent experience across the enterprise.
Build Governance That Scales
As AI adoption grows, governance becomes increasingly complex. Evaluating each solution with a different process creates unnecessary work for clinical, IT, compliance, legal and security teams.
Instead, organizations should establish a consistent governance framework with shared evaluation criteria, standardized documentation requirements, and common success metrics. A scalable governance model makes it easier to evaluate new technologies while ensuring they meet the organization's standards for safety, transparency, value and performance.
Evaluate Value, Not Just Capability
AI investments also need to make economic sense. A technology may perform well in isolation but still create limited enterprise value if it duplicates existing capabilities, adds significant implementation costs, or requires new infrastructure and support.
Organizations should evaluate not only what an AI solution can do, but what value it creates relative to its total cost. That includes direct financial impact, clinician time saved, operational capacity created, revenue improvement, implementation and integration costs, and whether the technology allows the organization to retire or consolidate other tools.
A mature enterprise AI strategy helps leaders make those tradeoffs consistently and direct investment toward the capabilities with the greatest clinical, operational, and financial return.
Measure Outcomes, Not Just Deployment
Implementing AI is only the beginning. Long-term success depends on clinician adoption.
Organizations that see the strongest results invest in physician champions, specialty-specific training, continuous optimization, and regular feedback loops. They also communicate openly about why AI decisions are made, how feedback shapes product improvements, and what clinicians can expect as solutions evolve. This transparency builds trust, encourages adoption, and reinforces that implementation is an ongoing partnership rather than a one-time deployment. Successful organizations recognize that implementation is a continuous process of learning, refinement, and improvement.
Technology alone doesn't transform workflows. People do.
Plan for What's Next
An enterprise AI strategy isn't just about solving today's problem—it's about creating a foundation for what's next.
Healthcare AI is evolving rapidly, and the solutions organizations choose today should be able to grow alongside their needs. Rather than evaluating a tool based solely on its current capabilities, leaders should consider how it fits into their long-term roadmap. Can it support additional workflows as AI matures? Will it scale across specialties and care settings? Does it simplify the technology landscape by reducing the number of vendors clinicians and IT teams need to manage?
The most successful AI investments aren't isolated point solutions. They're platforms that can evolve with the organization, support new capabilities over time, and reduce operational complexity.
Before selecting your next AI solution, take a step back and ask whether it strengthens the strategy you're building, rather than the problem you're solving today.
Putting Your AI Strategy into Practice
Before evaluating the next AI solution, healthcare leaders should consider:
- Does this support an enterprise priority or solve only a departmental need?
- How will it fit into existing clinical workflows?
- Will it simplify or increase operational complexity?
- Can it scale across specialties and care settings?
- Does it strengthen our governance strategy?
- How will success be measured one year from now?
Healthcare doesn't need fewer AI innovations—it needs more intentional AI strategies.
