How Bookmark Medical scaled clinical AI across 600 providers in value-based care

When Bookmark Medical set out to improve value-based care performance, the challenge wasn't a lack of patient data, it was too much disconnected information. Providers were spending valuable time piecing together patient histories instead of preparing for patient encounters. 

Two years after implementing Navina, more than 600 providers now use clinical AI to streamline chart review, support documentation, and coordinate care across 246,000 patient lives. 

At a July 2026 AMGA webinar, Bookmark Medical Chief Physician Executive David Hatfield, DO, MMM, joined Navina Medical Director of Healthcare Solutions Chaitanya Dahagam, MD, MS, to share Bookmark’s experience. 

In this article:

1. Why Bookmark Medical chose AI
2. How they earned clinician trust
3. What changed after implementation
4. Lessons for other healthcare organizations 

Why Bookmark Medical adopted clinical AI for VBC 

Bookmark Medical supports employed and affiliate primary care providers across Massachusetts, Michigan, Tennessee, and Arizona in value-based care arrangements with responsibility for total cost of care. 

Success under value-based care depends on having a complete, accurate understanding of every patient. That means documenting chronic conditions, identifying care gaps before the visit, and ensuring providers have the information they need to make the right decisions at the point of care. 

Before AI, most of that work was manual. 

Building a care plan meant scouring the chart for diagnosis codes, assembling a picture of the patient's disease burden, and translating the picture for the care team. "We did it by brute force," Dr. Hatfield said. Providers moved screen to screen through consultation notes and lab values for details supporting what they were already treating. 

Technology gave providers access to more data, but not more clarity. Instead, the growing volume of information was increasing administrative burden and eroding the joy of practicing medicine.

Bookmark Medical knew it needed a different approach.

The goal wasn't simply to improve documentation, it was to reduce provider burden while helping clinicians deliver better care.

That led Bookmark Medical to evaluate clinical AI that could simplify workflows, reduce cognitive burden, and support clinicians at the point of care. 

But Bookmark Medical wasn't looking for just another documentation or coding tool. Dr. Hatfield knew that if a new solution added clicks or disrupted clinical workflows, providers just wouldn't use it. 

Those requirements became Bookmark Medical’s benchmark for evaluating clinical AI solutions. 

Navina stood out because it could consolidate fragmented clinical information into a single patient view, surface evidence-backed insights, and integrate directly into the provider's existing workflow - all at the point of care. Rather than adding another application, it reduced the work required to understand the patient before the visit, helping providers spend less time searching through charts and more time caring for patients. 

Selecting the technology was only the first step. The next challenge was earning provider trust and adoption. 

How Bookmark Medical built clinician trust in AI

Bookmark Medical understood that selecting the right technology was only half the battle. For AI to succeed, providers had to believe it would make their work easier, and not introduce yet another system competing for their attention.

"You have to build a why or a culture into every workflow, or physicians and APPs are going to look at you cross-eyed like, 'You're just giving me another thing to do, another click, another [thing] to go look at. Get out of my way. Let me just see patients and take care of them.'"

Instead of asking clinicians to change the way they practiced medicine, Bookmark Medical positioned AI as a way to remove work from their day. Documentation, coding, and quality capture would happen within the normal flow of the visit rather than becoming additional administrative work after the patient left.

That message resonated. Adoption moved faster than Dr. Hatfield expected, and today Navina is embedded deeply enough into clinical workflows that a drop in HCC addressable rates is viewed as an operational anomaly worth investigating rather than a sign that clinicians aren't using the platform.

How clinical AI transformed workflows at Bookmark Medical

Once Navina became part of the clinical workflow, the impact extended beyond helping physicians prepare for visits. 

Providers spent less time piecing together patient histories, coders could focus their attention on more complex opportunities, and care teams gained better visibility into the patients who needed intervention. At the same time, keeping clinical evidence connected to every surfaced insight helped Bookmark maintain provider trust as AI became more embedded in care delivery.

Reducing chart review burden for providers and coders

Providers felt the change first. The assembly work now happens before the encounter, so clinicians walk in with the patient picture already built.

Coders "don't have to worry about the low-hanging fruit anymore," Dr. Hatfield said. A deeper pre-visit workup puts only genuinely uncertain items in front of a provider, leaving clinicians with a faster accept-or-reject decision. Post-visit cleanup has also shifted toward targeted education around the codes individual clinicians consistently miss.

Improving care coordination and quality gap closure

That same patient picture gives care teams a shared view of who is actually at risk. Bookmark Medical stratifies patients by RAF score and builds services around the highest-risk tiers, including care management, pharmacy, social work, and a visit cadence called a Stay Well care plan.

When a tier three or four patient misses an appointment, the outreach team treats the miss as urgent. "It's actually a medical emergency to say, 'Uh-oh, why did this patient miss their appointment?'" Dr. Hatfield said. Left unmanaged, those patients wind up in the ER.

Quality runs on the same point-of-care logic, dependent on claims data and on whether a screening done elsewhere gets recorded. Central teams still chase gaps from a distance, but the leverage sits in the exam room, and the physician doesn't have to be the one who acts: a medical assistant can place a mammogram order off a gap surfaced during the morning huddle. 

Many of Bookmark Medical's contracts carry quality gates tied to star levels, so point-of-care gap closure has direct financial weight.

Building clinician trust in AI-suggested conditions

For these workflows to work, clinicians have to trust the insights being surfaced by AI. 

Every condition Navina surfaces includes supporting evidence and a link back to the source document, whether that evidence comes from labs, medications, discharge summaries, imaging reports, or other connected clinical data. Clinicians can push a compliant note into the record, and anything already addressed during the visit isn't raised twice.

That evidence trail answers the objection physicians reliably raise about coding software. "Nobody's trying to practice medicine for me. Nobody's just trying to make me add codes to upcode the visit," Dr. Hatfield said of the reaction he hears. "They're just helping me paint the picture of the patient."

What other organizations can learn from Bookmark Medical's AI rollout

Dr. Hatfield has signed contracts with vendors that promised everything in the sales cycle and couldn't deliver. His recommendation: run a real RFP, then talk to people already using the technology. Ask about solutions for specific pain points like clinician adoption, burden reduction, HCC capture, and quality gap closure.. "Without data, you're just a person with another opinion," Dr. Hatfield said. "Healthcare does not need more opinions."

The criterion Dr. Hatfield weighs most heavily is whether a vendor treats a surfaced problem as its own. When Bookmark Medical's coders flagged gaps around provider education and newer codes like stage B heart failure, Navina worked with the Bookmark team to address them.

Bookmark Medical's experience shows that successful clinical AI adoption isn't simply about surfacing more information. It's about giving clinicians the right information at the right time, fitting it into the way they already work, and turning those insights into better documentation, more coordinated care, and less burden on providers. 

Watch the full session to hear Dr. Hatfield share Bookmark Medical’s experience firsthand, or book a demo to see how Navina can support your organization’s clinical workflows.