Medical Coding Tools Are Evolving Fast. Brad Watt Explains What Comes Next for Revenue Cycle

About This Blog

The journey into medical coding is rarely a straight line. Brad Watt, Chief Business Development Officer at InnoviHealth, started his career selling women's shoes with a freshly minted computer science degree. The year was 2004, the tech bubble had just burst, and commission checks from the shoe department were outpacing his opening tech salary offers. That unexpected detour led him through physician recruitment at Merri Hawkins, book sales at Optum, product management at Decision Health, and ultimately into building the AI-powered coding tools he works on today at InnoviHealth.

In this episode of the Fountain of Vitality podcast, host LaMont Leavitt talks to Brad Watt about the full arc of medical coding tools, starting with the printed manual era and ending with AI agents trained to know their own limitations.

 The Code Book Era and Why It Mattered   

Brad spent years at Optum selling code books, encoders, and training materials during a time when printed manuals were the backbone of every coder's desk. He recalls regional health systems spending a million dollars annually on code books alone. AAPC conferences featured Optum booths where dozens of coders played musical chairs just to win a single manual. Those books represented the accepted source of truth for the entire profession.

The coding manual still holds that status for many professionals today, particularly because certification exams require physical books. But Brad raises a point that challenges that assumption directly: coding manuals become outdated roughly 60 days after the effective date of the codes they contain. Regulatory changes, code updates, and even printing errors accumulate almost immediately. Errata documents get issued, but the book on the shelf stays the same.

LaMont added his own observation: many coders kept updated books on their shelves less for reference and more as protection during audits. Having a current edition visible signaled compliance, even if the actual reference happened elsewhere.

 The Transition to Online Tools   

The shift from print to digital didn't happen overnight. Brad describes the progression: first came digitized versions of the same information, then ebooks, then lookup tools that could pull code descriptors and basic guidance. Eventually, platforms like Find a Code and Select Coder began aggregating information from multiple sources into a single interface. Instead of flipping between an ICD-10 manual, a CPT book, a coding companion for a specific specialty, and a Medicare guidance document, coders could access everything in one place.

Brad's transition into product development at Decision Health was organic. Hired initially to sell books after they acquired Contexo Media, his computer science background pulled him toward the software side within six months. He spent 18 months diving into coder workflows, running weekly feedback sessions with customers, and redesigning how Select Coder displays information to match how coders actually think and search.

That experience taught him something he still applies: the relationship between online tools and the printed manual needs to stay close. Many coders still find their code in a physical book first, then enter it into software for additional context. The tool has to respect that workflow rather than fight it.

 The Expanding Role of Coders and Auditors   

Brad points out that the coder's job description has changed significantly. What started as reading documentation, assigning codes, and submitting claims now includes pre-submission auditing, post-denial investigation, and compliance analysis. He compares the modern auditor role to investigative reporting: digging into denied claims, figuring out what went wrong, and building the case for resubmission.

This expansion has created both opportunity and pressure. Coders face productivity KPIs that demand speed while handling increasingly complex work. Denied claims pile up, and many never get reworked simply because teams lack the time. Brad sees this as one of the most tangible areas where better tools can make an immediate difference. Faster, more accurate initial coding means cleaner claim submissions. That frees experienced professionals to focus on the complicated cases that require human judgment and persistence.

 AI as a Team Member, Not a Replacement   

Brad frames AI in healthcare coding as a coworker rather than a replacement. He compares it to working with a colleague named Bob: you'd rather have Bob say "I don't know" to a complicated question than deliver a confident answer that turns out to be wrong. The stakes in medical coding make this distinction especially important. A hallucinated code or fabricated guidance can trigger denied claims, compliance issues, and financial loss.

The approach at InnoviHealth involves training AI agents to search only within approved, accepted sources. When the AI cannot find a definitive answer, it's designed to acknowledge that gap rather than fill it with generated content. Brad believes this makes the tool trustworthy in a way that general-purpose AI often isn't for specialized professional use.

He also acknowledges the broader fear around AI in the coding profession. Computer-assisted coding triggered similar anxiety before ICD-10, with predictions that automation would eliminate coding jobs. Instead, the profession grew. More coders and auditors work in the field now than before those tools existed. Brad expects a similar pattern with AI: the tools will accelerate the work, not replace the workers.

 The Right Tool for the Right Job   

A recurring theme throughout the conversation is Brad's belief that success depends on matching the right tool to the specific task. He traces this philosophy back to growing up on an Idaho farm where nothing got fixed without the correct wrench, the right attachment, or the proper equipment for the job.

He applies the same thinking to Find a Code and the broader product ecosystem at InnoviHealth. The goal isn't to build one tool that does everything. It's to build a toolbox where each component solves a specific problem well. Online coding tools, AI agents, auditing resources, and educational materials all serve different functions for different moments in a coder's workflow.

 Key Takeaways   

1. Medical coding manuals become outdated within approximately 60 days of the code's effective date due to regulatory changes, updates, and printing errors

2. The coder's role has expanded well beyond code assignment into auditing, compliance, investigation, and pre-submission quality review

3. AI tools trained on approved sources with built-in uncertainty acknowledgment offer more value than general AI that fills knowledge gaps with generated content

4. The profession has consistently grown alongside automation rather than shrinking, from computer-assisted coding through the current AI era

5. Tool design should follow the way coders actually work and think, rather than forcing new workflows that abandon familiar processes

Step inside the world of code books, conference floors, and AI agents that know their limits with Brad Watt on the Fountain of Vitality podcast. Explore how coding tools evolved from printed manuals to intelligent systems and what that means for revenue cycle professionals chasing cleaner claims and faster reimbursement.

Follow the Fountain of Vitality podcast: Website: FountainofVitality.com | Tiktok: @FountainofVitalitypod | YouTube: @FountainofVitalityshow | Tumblr: @FountainofVitality | Facebook: FountainofVitalityShow | Rumble: Fountain_of_Vitality
| Instagram: @FountainofVitalitypodcast | Email: contact@FountainofVitality.com

Follow LaMont Leavitt: LinkedIn: @LaMontJLeavitt/ | Twitter/X: @ljleavitt1 |
InnoviHealth Website: InnoviHealth.com

Follow Brad Watt: LinkedIn: @Brad-Watt-90203527