Readers Write: The Future State of AI and Automation in the Revenue Cycle

The Future State of AI and Automation in the Revenue Cycle
By Patrice Wolfe

Patrice Wolfe, MBA is CEO of AGS Health.

Like many heavily regulated industries, healthcare has seen limited progress towards the use of artificial intelligence (AI) and automation, despite the enormous potential they hold for improving productivity, accuracy, care access, and the bottom line. Much of that promise comes from use cases that span the revenue cycle management (RCM) continuum, where legacy automation tools are already having a positive impact through activities like patient reminders, insurance verification, coding, and claims status transactions.

Today, generative AI (GenAI) is poised to upend, in a positive way, healthcare’s approach to front- and back-end financial operations. It has the potential to re-imagine the massive volumes of historical and real-time revenue-related data that is flowing through RCM departments and create entirely new approaches to optimize revenue and minimize financial risk.

AI is an advanced set of tools run by algorithms that use data to simulate human intelligence. GenAI takes things several steps further by leveraging that same data to not only tell the story of what it sees, but to also create entirely new, more effective approaches to RCM.

Already present in many RCM functions, AI and automation represent a continuum of capabilities that can be broken down into four major categories:

• Basic. Rules-based processes for repetitive tasks that typically follow pre-defined instructions without exceptions. Examples include a claim status or transaction query submitted by a provider using a basic bot or ANSI transaction that returns a response based on a predefined set of values.
• Advanced. Leverages more complex algorithms and machine learning to make predictions based on past performance, which allows for proactive intervention based on those probabilities. For example, a machine learning model may be able to identify claims that are likely to be denied and can be corrected before being submitted to the payer.
• Intelligent. Here is where AI enters the continuum with the addition of natural language processing (NLP) that uses unstructured data and human-like reasoning to process ambiguity. An example in RCM would be the use of machine learning, deep learning, and NLP models that recommend “next best actions” to prevent denials from even happening in the first place.
• GenAI. Uses neural networking and large language models (LLMs) with deep learning and other techniques to automate design and do complex problem solving, often aided by visual and written materials. An example would be a human-like chatbot that negotiates with payers to reverse claim denials using the clinician’s notes and imaging studies to develop an argument complete with appropriate medical terminology.

While healthcare remains in the early stages of the AI continuum, more complex and sophisticated Intelligent and GenAI use cases are on the horizon.

While all eyes are on GenAI, earlier-stage AI and automation is already impacting RCM outcomes and efficiencies. Meanwhile, ample opportunity exists to further influence RCM as capabilities grow. In fact, just as AI and automation fall on a continuum, so too do the RCM processes and workflows that can be boosted by their adoption.

Scheduling and Registration

Legacy automation has a stronghold in scheduling and registration with the use of basics like automated patient reminders now nearly ubiquitous among healthcare organizations. Looking toward the future, scheduling chat bots, integrated scheduling across care sites and clinical specialties, and comprehensive scheduling packages for patients that include cost estimates are high-priority investments for their potential to reduce patient friction, enhance the patient experience, and make a provider “stickier” by strengthening the provider-patient bond and improving patient retention.

Patient Access

Insurance and benefit verification are already close to fully automated. RCM’s holy grail of future automation use cases is prior authorization, particularly as payers build more complex and ever-changing policy requirements for prior auth. AI can help manage the prior auth process, maximize the probability of approval, and automate the appeals process if an authorization is denied. The challenge is the enormous amount of information that is required from both providers and payers who have little incentive to be transparent with those details.

Coding/HIM

Computer-assisted coding (CAC) enjoys broad adoption for inpatient coding and billing, delivering reported productivity gains of 10% to 30% for hospitals. Computer-assisted professional coding (CAPC) is beginning to make inroads on the professional side. Future use cases include autonomous coding, which has limited use in a handful of specialties due to the significant amounts of data needed to properly train the specialty-specific LLMs. Early work is also underway around ambient charting, which converts voice dictation into coding and promises to save physicians up to 4.5 minutes per chart by some estimates.

Patient Financial Services

As with prior authorization, AI and automation adoption in patient financial services is influenced by increasingly aggressive payer policies around denials, delays, and underpayments. There is enormous potential for streamlining collection workflows, including touchless A/R. Other promising areas are automated denials management and the movement to a reduced friction patient experience.

Clinical Services

Though farther behind other stops on the RCM continuum, future AI and automation use cases within clinical services include real-time patient status monitoring in utilization management (UM) to ensure accurate reimbursement. Other potential applications include professional fee UM and automated clinical documentation integrity (CDI) that uses NLP and other advanced tools.

Revenue Integrity

Also behind the adoption curve, revenue integrity AI and automation use cases include charge master maintenance, late charge identification, and coding/billing compliance audits. AI and automation are also used to proactively identify and resolve problem areas.

Healthcare has taken a cautious approach to adopting GenAI and other advanced forms of AI and automation within RCM, due in part to the industry’s necessarily risk-averse nature. Also at play are the complexities that are involved with adapting critical workflows to advanced AI and the need to balance the application of limited resources between multiple and sometimes conflicting strategic priorities.

For example, while advancements like ambient documentation are crowd pleasers that promise to deliver improvements in physician productivity and satisfaction, they won’t necessarily improve the completeness of clinical documentation. As such, CDI will remain a critical part of the RCM process.

The reality is that while GenAI and its AI peers hold great promise for optimizing RCM, these technologies can be expensive to use, staff, and support. Health systems and other provider organizations will have to place bets with scarce resources, and it’s more likely that AI use cases that improve physician and patient satisfaction will come out on top.

GenAI and advanced automation also require close collaboration between operating departments like RCM and their IT colleagues to create and test APIs, move/share data between systems, and access datasets to test predictive models and train LLMs and other advanced AI models. This collaboration may be hampered by information and data silos that were created by legacy technologies. This also impacts the opportunity to leverage AI and automation to create a seamless patient experience, which requires integration across multiple settings of care, systems of record, and data siloes.

As GenAI and other advanced automation solutions continue to deliver on their promise, the impact on healthcare RCM has the potential to be transformational. They also have the potential to reduce the challenges that are confronting providers across the RCM continuum, while streamlining patient access, increasing coding and billing accuracy, improving utilization management, and speeding the revenue cycle.

When the productivity and accuracy promises are fully realized, investing in GenAI and its predecessors becomes a true win for the entire healthcare industry.