RPA vs. AI for medical billing: What does the future hold?

In times of increased administrative burden and healthcare staffing shortages, there’s an important question on every revenue cycle manager’s mind: Will artificial intelligence — otherwise known as ‘AI’ — take over medical billing? Imagine if you could double the healthcare services you provide without hiring any additional revenue cycle staff to submit necessary healthcare claims — all while avoiding denials and improving cash flow. With AI, this dream could become a reality.

Some industry experts estimate the wider adoption of AI could reduce United States healthcare spending by up to $360 billion annually. An added bonus? It could also improve patient satisfaction. 

When AI handles less complicated, repetitive tasks, this frees up humans (i.e., revenue cycle management [RCM] staff) to focus on things like answering detailed coverage questions, explaining healthcare costs, or simply helping patients physically navigate to the waiting area — all of which enhance the patient experience. 

Although some may have concerns that AI may replace RCM roles, AI ultimately supports human work in this area. It does not eliminate it. 

While no one has a crystal ball, exploring the potential of AI — especially in the healthcare revenue cycle — makes sense. In this article, we’ll explore key terminology and discuss the impact of automation, machine learning, and other forms of AI on healthcare RCM. 

The healthcare industry — and specifically RCM — has reached a critical tipping point in which healthcare organizations can’t keep up with the sheer volume of claims. Thirty-two percent of US adults avoided routine care during the years-long pandemic, and the consequences of those delays are only now coming to light through pent-up demand for healthcare services. In addition, aging populations and longer life expectancies continue to increase claim volumes as older individuals tend to use more healthcare. 

As the demand for healthcare increases, there are fewer and fewer revenue cycle staff to collect and submit critical information to payers. Thirty-four percent of medical groups find medical coders the most difficult revenue cycle role to hire, according to a recent survey. 

Those in the medical coding and billing field also require specialized education and training and are generally more expensive to hire than other revenue cycle staff. As a result, providers increasingly look to automation and AI tools to help existing staff members gain efficiency and minimize the need to hire additional staff. 

With the right artificial intelligence privacy and security policies, procedures, training, and change management strategy, the technology can also help mitigate the alarming problem of revenue cycle staff burnout. For example, AI removes manual, repetitive tasks for front-desk staff so they can focus on more meaningful and satisfying work, such as building rapport with patients. Similarly, AI gives medical coding and billing staff the confidence they need to select the right codes, thereby removing the guesswork and stress associated with non-compliance.  

RPA is the acronym for “robotic process automation” — the ability to leverage intelligent “bots’’ to perform manual, repetitive functions — and in the healthcare revenue cycle, there are many of them. 

For example, RPA can check payer portals for claim payment status, thereby eliminating the need for billing staff to spend countless hours on the phone or on payer portals to obtain updated information. RPA can automate responding to payer requests for additional documentation. 

It can also automate verifying patient insurance, charge entry, validating and downloading/uploading electronic remittance advice and explanation of benefits files from payer portals to the practice management system, posting and adjusting payments, and much more. 

Many RCM tasks are ripe for robotic process automation. By automating repetitive tasks, such as data entry, scheduling, billing, and more, healthcare providers can reduce errors and improve patient satisfaction. 

AI for RCM is an umbrella term for technology that performs complex and cognitive functions based on its ability to learn from the data — not simply replicate and automate rule-based tasks. Think of AI technology as a “step up” from RPA — in other words, a more advanced form. While RPA focuses on automation, AI introduces an element of human intelligence (e.g., reasoning, learning, problem-solving, and decision-making).

Examples of AI technologies include: 

Like robotic process automation, each of these technologies has a place in healthcare RCM. For example, predictive analytics can predict what claims payers may deny and why, providing valuable insights to prevent denials at the front end of the revenue cycle and reduce rework. 

With a combination of machine learning and natural language processing, you can automate certain aspects of medical billing and coding. Together, these technologies (collectively referred to as computer-assisted coding technologies) can analyze clinical narratives, physician notes, and other unstructured text data to extract relevant medical codes and information with incredible precision. With generative AI, you can create original, fact-based appeals to health insurers. 

It’s not an “either” “or” scenario. Both RPA and AI have a place in healthcare, and both can help address specific revenue cycle pain points and challenges. 

For example, if you struggle to retain enough billing staff to post payments in a timely manner, RPA can be extremely advantageous in terms of keeping up with workload demands. However, if you struggle with countless denials that payers typically overturn on first-level appeals, a generative AI tool can help create appeals. This allows staff to focus on more complex appeal letters. Generative AI would rely on past appeal letters, payer policy manuals, and contracted terms to write unique letters for each case. 

Also, keep in mind that some technologies inherently include a combination of RPA and one or more types of AI, whereas others include strictly one or the other. As you consider prospective vendors, it’s important to know what technology is — and isn’t — embedded into their product and how it may — or may not — solve the specific RCM challenges you face.

AI and RPA are considered complementary technologies. When leveraged together, medical practices can achieve what some experts refer to as intelligent automation, or IA. 

IA is the winning combination that, when augmented by human oversight, has the potential to enable widespread digital transformation. In healthcare RCM, this digital transformation can lead to increased revenue cycle productivity, reduced costs, improved billing accuracy, and a better overall patient experience. 

With AI and IA, human oversight is critical. That’s because these technologies supplement or augment a human workforce and provide opportunities for upskilling, retraining, and professional growth. 

The key to implementing AI in the revenue cycle is that technology must complement the workload and workflow of coding, billing, and administrative professionals. Revenue cycle staff and technology must develop a synergy in which staff verify the output from the AI solution and continually provide a feedback loop for improvement and validation. Ultimately, AI supports human work. It does not eliminate it. 

The more important question is this: Will AI eventually become a “need to have” rather than a “nice to have” aspect of RCM? The answer is yes. 

Organizations using manual processes will not be able to keep up with the pace of claim volumes coupled with revenue cycle staff shortages and labor costs. Antiquated operational models simply will not be sustainable, and they will undoubtedly cause denials, cashflow problems, and other RCM challenges. 

Just as providers continue to explore the use of AI in RCM, payers are doing the same. For example, many payers use algorithms and AI-based tools to determine coverage limits, make claim determinations, and engage in benefit design. 

The Centers for Medicare & Medicaid Services has focused specifically on Medicare Advantage plans, stating these plans are responsible for ensuring that an algorithm or an AI-based tool complies with the agency’s coverage decision requirements. 

In addition, payers cannot use AI to shift coverage criteria over time, nor can they apply predictive algorithms and internal coverage criteria that are not public. This guidance comes in the wake of several lawsuits against payers (i.e., UnitedHealthcare, Humana, and Cigna) allegedly using AI tools or algorithms to wrongfully deny care to Medicare Advantage members. 

The takeaway point for healthcare providers? Pay attention. Payers’ use of AI may go completely under the radar, and it’s important to leverage equally powerful tools to protect medical practice revenue. 

Medical practices continue to embrace automation and AI to optimize billing processes, detect and prevent fraud, and mitigate the risk of denied claims. AI and machine learning applications can improve revenue cycle efficiency while simultaneously enhancing the patient experience. 

In the future, AI will evolve to provide more personalized and patient-centric billing solutions, integrate seamlessly with electronic health records, reduce errors and improve transparency in billing, and enable greater customization to solve specific needs and challenges. 

Remember, human oversight is critical when it comes to using AI. Verifying outputs from AI solutions and continually working to improve processes will be key to getting the best results.

AI is poised to become an integral part of revenue cycle management because of its ability to process data, provide insightful analytics, and drive efficiency and accuracy. Medical practices that leverage it will establish a foundation for financial sustainability. 

Find out how your practice can dramatically increase productivity and profitability with Tebra’s intelligent robotic process automation solution.