ChatGPT And GenAI In Medical Billing: What You Need To Know
Is artificial intelligence (AI) taking over medical billing (aka revenue cycle management / RCM)? Is this something you need to worry about right now, or something you should dedicate time and resources to implementing? There’s a lot of discussion around AI, especially generative AI like ChatGPT, which is receiving the most attention.
Recently, I had a conversation with a multi-billion-dollar private equity group looking to acquire companies in the RCM space. This is just one of many similar discussions I’ve had over the past year. These conversations often center on the impact of AI in medical billing companies, particularly generative AI.
AI in Medical Billing and RCM: Deals and Investments
When I see RCM businesses coming to market—especially those represented by large investment banks like Guggenheim or Houlihan Lokey, with enterprise values ranging from a quarter-billion to multi-billion dollars—there’s almost always something about artificial intelligence in the marketing materials. Typically, there’s more hype than substance. It feels necessary to include AI in medical billing to sell the business at a high valuation these days, and I believe this trend is somewhat overblown, particularly when it comes to AI in medical billing.
Here’s my position: while generative AI is hot and incredibly promising for marketing and technology when it comes to RCM, the situation is more complex. Should you worry about AI in RCM right now? It depends on whether you’re looking at the short term or the long term.
The Long-Term and Short-Term Impact of Generative AI in Medical Billing and RCM
Long-term, there will undoubtedly be a massive impact across industries, especially in areas like AI in medical billing. It may even become integral to the overall (or certain sub-components of) workflow within revenue cycle management. However, in the short term, using LLMs like ChatGPT shouldn’t be your primary focus. Even if your organization has already perfected other aspects of its technology platform, the benefits of generative AI have more potential than proven right now.
To be clear, when I refer to AI, I’m not talking about automation, analytics, or predictive analytics using ML. I’m specifically referring to large language model (LLM) artificial intelligence, also known as generative AI, which currently has extremely limited (or no) impact in RCM.
Evaluating the Real Benefits of AI in RCM
There are undoubtedly some valuable applications of AI that can benefit RCM, but generative AI like ChatGPT is less useful for many of the areas often suggested within RCMN. Let me give you an example: one revenue cycle management company I evaluated had a valuation north of half a billion dollars. In their materials, they listed the potential benefits of generative AI for tasks like pre-authorizations or generating appeal letters.
When asked about this by a potential acquirer, my response was, “What makes for a good pre-authorization or appeal letter?” The answer lies primarily in understanding the payer policies – and ensuring compliance with those rules. Additionally, it’s crucial to know what logic or documentation supports an appeal. These are the critical factors in winning an appeal, rather than whether a human or a computer writes the text.
The value lies in the data, logic, and payer policy compliance, not in the automatic generation of text where the words are frequently paired near each other. Compared to copying and pasting compelling appeal information into a template, I doubt generative AI would offer a significant productivity boost over manual methods.
Conversational AI
Another type of artificial intelligence is conversational AI, which comprises AI for voice communication. This is typically used to communicate with insurance payers to check claims status, check eligibility and benefits, and to get prior authorizations completed. This technology can have extremely strong benefits and offer great ROI. While technically it is a form of generative AI, conversational AI is not really in the written LLM model space like ChatGPT, Grok, or Claude and so I would consider it a separate category.
Future of AI in RCM
There may be future applications of generative AI where language models are used to process words and data in revenue cycle management. However, in the short term, whether there’s significant value in written generative AI for pre-auths or appeals is highly debatable. I have yet to see anything that really impresses me. These tools may become valuable in the future, but for the short term, there are other areas where you should focus your resources.
Where to Focus in Medical Billing Technology
Instead of obsessing about generative AI like ChatGPT, it may be better to invest in proven technologies that deliver real value in the short term. These include solutions that enhance profitability, improve marketing and sales differentiation, boost client retention, and produce better overall results for your medical billing business.
For example, if you haven’t already implemented centralized analytics across multiple platforms, it’s probably premature to consider generative AI. You need to be able to quantify the benefits of any new technology you implement, regardless of what it is. Make sure your foundational technology is in place and fully optimized before exploring newer, untested AI applications.
In conclusion, while generative AI may eventually transform RCM, there’s no need to rush into generative AI just yet. Direct your attention to conversational AI or focus on the fundamentals that will have the greatest immediate impact on your business like data infrastructure.
