How AI Is Solving Dental RCM's Complexity Problem

Revenue cycle management (RCM) doesn't get talked about the way clinical AI or patient experience does, but according to Teresa Williams, CFO and COO of Dental Express, it might be the hardest operational problem in dentistry to actually solve. She joined Alan Rencher, CTO of Henry Schein One, on a recent episode of Dental TeX-ray to talk through why RCM is so complex, where AI is already changing it, and what's still missing even with better tools in place. 

Here are the key takeaways. 

RCM's complexity is, quite literally, too much for a human to fully hold at once 

Rencher framed the scale of the problem using a concept borrowed from software engineering: cyclomatic complexity, a score used to measure how many possible paths exist through a system. He noted that a score of 100 is considered effectively incomprehensible for a person to fully track, and that dental RCM in the US, with roughly 1,800 payors each applying rules differently, scores somewhere around 200. 

Williams didn't disagree, she reframed it from the operator's seat: "I would say it could be true, not at the same time. Eventually I would say that a human could eventually understand it all, but not all at the same time." 

Manual RCM created real, ongoing revenue leakage 

That complexity isn't theoretical. Williams described a persistent, quiet cost that many smaller practices simply absorbed as normal. 

"We had a lot of leakage in our industry. There was opportunity because there was so much leakage, because if you were dependent upon someone who was very charismatic with patients, they may not be onto those RCM things... they were going to miss a waiting period or miss a missing tooth clause." 

That leakage compounds at scale. A missed waiting period or an overlooked exclusion clause might be a rounding error for one practice; across a multi-location DSO, it becomes a material hit to cash flow. 

Embedded AI eligibility tools beat outside vendors on cost, speed, and accuracy 

Williams described Dental Express's own evolution: starting with an outside AI-powered eligibility vendor, then switching once eligibility checking became embedded directly inside their practice management platform (PMS). 

"AI embedded within our practice management software just really caught up... it was a huge cost savings when that happened, but it was a huge time savings. Things were more accurate, because even when you're using an outside vendor, if they're not in your business every day and understanding your workflows, there's still going to be some things where your end user says, 'Well, that just doesn't work.'" 

Her practical framing of "good enough" was refreshingly candid: a tool that works 98% of the time, freeing staff from sitting on hold with insurance portals, is a clear win even without chasing perfection. 

Real-time eligibility checking has to happen before the appointment, not after 

One of the sharper points in the conversation: knowing a patient's eligibility isn't useful if that information doesn't reach the clinical workflow at the right moment. 

"It's great that you found out I have a waiting period on SRP... but it's also put the percentage into my PMS. My hygienist goes to perform the care thinking it's covered at 80% when it's really covered at zero. How do I stop her?" 

Williams framed this as the next frontier for RCM technology: not just gathering better eligibility data, but building real interoperability between RCM systems and clinical workflow, so that a coverage gap actually creates a stop or a flag before treatment happens, not a surprise on the bill afterward. 

RCM failures aren't just financial, they damage the patient relationship 

Rencher shared a personal example: his daughter's recent dental procedure went smoothly precisely because RCM was invisible, eligibility had already been resolved ahead of time, so the clinical visit could focus entirely on her care and comfort. 

Williams connected that directly to patient retention. "It really sours the experience... we cannot support our clinicians to have these beautiful interactions and relationships with patients if their patients are counting on us... they're counting on us to be the experts." Practices that treat insurance complexity as someone else's problem, in her view, are cutting themselves off from real relationships with patients who are depending on them to get it right. 

The hardest people to bring along on RCM technology change are often the most experienced 

Asked about adoption challenges, Williams offered an unusually self-aware answer: the people most resistant to new RCM tools tend to be the ones who already mastered the old, manual complexity. 

"If those are your decision makers in RCM, people that already understand the complexities of these things and they've designed a system around it, they are the hardest nuts to crack. I was one of those people... you have to realize you might be the person that's in the way because you do understand it all." 

Her fix was a personal habit rather than a mandate: deliberately reviewing every new feature or prompt a system surfaces, rather than staying on the autopilot clicks that already feel comfortable. 

The goal isn't more dashboards, it's a shorter list pointed at real dollars 

Asked for a closing takeaway, Williams reframed what "AI in RCM" should actually deliver for an operator. 

"It's not glamorous when you talk about I'm going to use AI on my stack with eligibility, your clean claims, your ERA matching... The glamorous part for me is if you can solve those problems with the bots, you get those things taken care of and serve up a very targeted list to your super users... here's where you go get the money and the right money you should be going after." 

The bottom line 

RCM's complexity isn't going away, 1,800 payers with different rules guarantees that. But the conversation with Williams pointed to a clear shift: AI embedded directly inside the practice management platform is proving more accurate and cost-effective than bolted-on third-party tools, and the next real gain comes from connecting that data to the clinical workflow in real time, not just collecting it faster. Getting there also requires being honest that the biggest obstacle to adoption is sometimes the most experienced person in the room, not the newest hire.

Check out the full episode now wherever you listen to podcasts.