Clinical documentation improvement has always existed in the space between patient care and reimbursement. It’s the bridge connecting what physicians do clinically with what organizations are ultimately able to code, bill, and collect. The problem is that traditional CDI processes have historically operated too late in the cycle — after the note is signed, after coding review begins, and often after valuable clinical specificity has already been lost.
That’s the challenge Zotec Partners plans to tackle head-on during HFMA Annual 2026 in National Harbor, Maryland, with a live presentation that moves the AI conversation beyond hype and into practical application.
On Tuesday, June 9 from 4:10 PM to 5:00 PM, Jimmy Comfort, VP of AI R&D at Zotec Partners, will present From Dictation to Dollars: A Beginner’s Guide to Building Real-Time CDI with AI. The session explores what happens when clinical documentation improvement no longer works retrospectively, but instead operates in real time — inside the physician workflow while documentation is still being written.
For years, CDI teams have worked reactively. A physician signs a note, coding reviews it later, and eventually a query is issued to clarify missing documentation. The process creates friction for physicians, delays for coders, and rework across the revenue cycle. In many cases, opportunities are missed entirely because the clinical intent was never fully documented in the first place.
Real-time AI-powered CDI changes that equation.
Instead of relying on retrospective review alone, context-aware large language models can now analyze an encounter while documentation is actively being created. That includes reasoning across physician notes, nurse notes, medications, procedures, lab results, and orders simultaneously to identify documentation gaps before the chart is finalized. In practice, the experience feels less like an audit and more like spellcheck for clinical documentation.
The timing is what makes the difference.
Rather than surfacing issues days later through queries and addenda, AI-generated CDI nudges can prompt physicians in the moment with compliant, encounter-specific documentation suggestions. During the HFMA session, attendees will see real-world examples involving EKG interpretation documentation and critical care attestations — two areas where documentation gaps frequently impact reimbursement accuracy and RVU capture.
One example demonstrated in the presentation involves an emergency department physician documenting that an EKG was “normal.” While clinically understandable, that language alone may not support a fully compliant interpretation. The AI identifies the missing specificity and immediately suggests language such as rhythm interpretation, heart rate, and absence of ischemic changes, allowing the physician to quickly complete the documentation before signing the note. The result is less downstream rework and more accurate capture of the care already delivered.
The financial implications are significant. During a 1,000-physician pilot referenced in the presentation, organizations experienced an average RVU lift of 3.3%, with top performers improving by more than 13%. More importantly, the pilot demonstrated that over half of participating physicians improved documentation quality without adding new workflows or requiring additional staffing resources.
That operational reality matters in today’s healthcare environment. Health systems and physician groups are already balancing workforce shortages, increasing regulatory scrutiny, reimbursement pressure, and growing clinician burnout. The last thing providers want is another disruptive workflow layered into the EHR experience. Real-time CDI only works if it feels supportive rather than intrusive.
That’s one of the more compelling aspects of Zotec’s approach. The session focuses heavily on practical usability and physician adoption rather than abstract AI theory. The presentation breaks prompt-building into understandable components — role, context, task, and output format — showing attendees how CDI logic can be created using plain English instructions rather than complicated technical development. In fact, one of the core messages throughout the session is that if someone can clearly describe a documentation gap, they can likely help build the prompt needed to identify it in real time.
The presentation also emphasizes that these models improve continuously through physician interaction. Accepted nudges reinforce successful documentation patterns, while dismissed suggestions help refine the model and reduce unnecessary interruptions over time. That feedback loop is critical because successful AI-driven CDI depends on trust, accuracy, and minimizing alert fatigue.
Perhaps most importantly, the session positions AI-powered CDI not as a future-state concept, but as something organizations can begin implementing today. The technology already exists to bridge the longstanding gap between clinical documentation and reimbursement — the industry is simply learning how to operationalize it effectively.
For organizations focused on improving coding quality, reducing rework, strengthening physician documentation, or optimizing revenue integrity, this session offers a practical look at where CDI is headed next.
HFMA attendees can learn more during the session or visit Zotec Partners at Booth #272 throughout the conference in National Harbor.
From Dictation to Dollars: A Beginner’s Guide to Building Real-Time CDI with AI
📅 Tuesday, June 9, 2026
🕓 4:10 PM – 5:00 PM
📍 HFMA Annual Conference — National Harbor, MD
📍 Booth #272
