As artificial intelligence (AI) continues to revolutionize industries across the world, healthcare revenue cycle management (RCM) stands out as a field ripe with opportunity. From automating administrative tasks to predicting denials trends, AI integration is no longer just a possibility but a necessity. Here are five things to think about when considering how AI can fit into your RCM and denials management strategy to optimize medical claims reimbursements.
1. Diverse Applications for AI Across the Revenue Cycle
AI’s broad yet intricate impact on healthcare RCM extends across the entire revenue cycle. It is being used in myriad ways, from automating data capture during clinical interactions, to improving coding and customer service operations, to appealing denials. After all, almost half of healthcare providers in the US have witnessed a surge in claim denials, with errors in patient access and registration emerging as the primary culprit. These results highlight the relentless hurdles providers encounter in securing reimbursement. By automating routine tasks, healthcare providers can focus more on patient care and less on administrative burdens.
2. Strategic AI Implementation Is Crucial
A structured approach is needed for AI implementation, including defining automation types, data application, and subscription models, as well as understanding how machine learning (ML) and large language models (LLMs) fit into the structure.
“Machine learning is a broad category, and large language models are the newest capability within that broad category,” says Aspirion Chief AI Officer Spencer Allee. “The more traditional machine learning is going to be things like regression or classification, where you’re taking some data and you’re predicting kind of a discrete outcome. Large language models are really focused more on natural language, either extracting information from that natural language or generating text—think ChatGPT.”
Jim Bohnsack, Aspirion’s Chief Strategy Officer, says Aspirion is using ML predictive models to prioritize which accounts are most likely to be overturned or resolved. “Using those models to say, what are the characteristics of this account?” says Bohnsack. “If I have a hundred accounts, in which order do I work those hundred accounts? Highest priority. Those are the traditional machine learning models.
“The newer large language models are able to take unstructured text loaded into these large language models and query it or pull out pertinent information, understanding the relevance of the words within the text to do some kind of generative output. For example, create an appeal letter based on the evidence of the medical record, which is one of the use cases we’re doing now.”
Successful AI implementation requires a clear understanding of the types of automation applied, the data used (resting vs. live data), and the maintenance necessary post-deployment. Providers must evaluate their core competencies, system stability, and long-term maintenance capabilities to decide whether to build internally, partner with external vendors, or purchase solutions.
3. Managing Payer-Provider Relationships with AI
The relationship between payers and providers is becoming more complex, contentious, and critical, especially with the rise of Medicare Advantage and the need for robust contract management and payer monitoring. Some payers are more stringent in their interpretation of contracts and rules, making it essential for providers to understand and manage their contracts meticulously. Breaking down walls between payer and provider data can improve member experience and reduce friction in care delivery.
AI can help bridge gaps by automating processes and improving data accuracy, but integrating payer and provider systems remains challenging, especially when, according to the American Medical Association (AMA), approximately 11 percent of all claims were denied by payers in 2023, up from 8 percent in 2021. That 11% rate translates into 110,000 unpaid claims for an average-sized health system and billions of dollars in lost revenue annually. The increasing complexity of billing and coding, along with frequent changes in payer policies—which in some cases call for a blanket denial-first policy—contributes to the rising number of denials.
4. Navigating the Future of AI and Provider Labor Shortages
The adoption of AI in healthcare RCM is accelerating, driven by the need to address labor shortages and rising costs. An analysis of 40 interviews conducted with healthcare executives revealed that 63% of providers were facing staffing shortages in their RCM departments. Providers were prioritizing the hiring of RCM talent to enhance the patient experience; however, only 42% of providers reported satisfaction with their patient pay solutions.
“How do you help speed along the process through which a coder or a clinician, in our case, an attorney, reviews up to a 2,000-page medical record and pulls out the pertinent information to justify back to the payer why it was that complex of a case or should have been paid inpatient?” says Bohnsack. “The DRG shouldn’t have been downgraded. Those types of things are being enabled through large language model usage today which is a step change in technical capabilities and is evolving really fast.”
Hospitals must be cautious and strategic in their approach, ensuring that AI applications are well-defined, thoughtfully applied, and properly maintained to avoid potential pitfalls and maximize benefits. The pressure from labor constraints makes AI an attractive solution for maintaining efficiency and consistency.
5. Evaluating Vendor Partnerships
Whether to build, partner, or buy AI solutions is a critical consideration for healthcare organizations. The decision-making process involves making evaluations based on core competencies, environmental stability, and specific needs. External vendors like Aspirion offer AI capabilities either as a service or as a product to be integrated into workflows, which are two distinct approaches. Vendor partners can offer specialized and quicker solutions, but internal teams may provide more tailored implementations.
The decision should be based on the provider’s competencies, system stability, and long-term maintenance needs. A thoughtful approach can lead to more effective AI integration, ensuring hospitals and health systems can leverage AI’s full potential without compromising quality or efficiency.
Unlocking the Full Potential of AI in Healthcare RCM
AI holds immense promise for hospital RCM, offering solutions that streamline operations, improve data accuracy, and maximize denials prevention and resolution. By approaching AI implementation strategically and thoughtfully, hospitals and healthcare systems can unlock its full potential and navigate the future with confidence.
For more insights into specific uses of AI in the revenue cycle, check out this on-demand webinar, “AI Results and the Provider Perspective.”
Ready to begin your journey toward optimal AI-accelerated RCM and increased profit margins? Contact us today.
