With the recent controversy over generative AI in Hollywood (likeness theft), academia (cheating), and business (chatbots), you may be concerned about how using this technology to generate information will affect the healthcare revenue cycle. However, even though generative AIs present a novel use for our industry, ChatGPT is used by Epic this way with some health record processing tasks.

On Google Cloud, AI-augmented prior authorization processing is now available. Telemedicine platform Doximity is rolling out a ChatGPT application that can generate preauthorization and appeal letters. Currently, there are barely any live examples of generative AI at work in a healthcare organization. A New Orleans-based healthcare system is using ChatGPT to assist clinicians in responding to emails with its email-based system. Chapel Hill's UNC Health has also launched an internal chatbot that is generative AI-type and powered by the Microsoft Azure OpenAI Service to answer queries, provide instant recommendations for solutions, and instruct visitors on destinations.

This is a controversial update to the traditional AI model and now boasts allegedly generative properties in the healthcare revenue cycle. Have no fear; there is more than enough time to look at what these exciting and a bit frightening new technologies may offer.

Understanding AI's Role in Revenue Cycle Management
To find the right answer for your revenue cycle, you need to understand at a deeper level what AI is and does, interpreting both its possibilities and perils. What follows is a head-to-head look at the improvements that each can bring to your revenue cycle if implemented correctly.

AI in RCM: Traditional vs. Generative AI
Before using generative AI—immediately garnering significant attention—a revenue cycle manager must realize that current “AI-driven” revenue cycle solutions used by most health systems are powered by traditional, not generative, AI.

Old-school AI, or as it's more concretely referred to these days, "machine learning," powers those eligibility checks, patient estimates, and prior auth headaches you've got; it also runs your account recovery pieces. It's an example of what we can call problem-solving AI, which uses large data sets to find patterns and do very specific tasks. It can make predictions or take actions that are learned to be done without being programmed explicitly, but the algorithm in its criteria is based on human-defined rules. Sometimes veiled patterns cannot even be caught by experts, but traditional AI allows for catching them.

Traditional AI Applications in RCM
Check Patient Eligibility: AI algorithms can enable systems to quickly check patients' eligibility, benefits, and coverage, leading to reduced servicing costs for ineligible patients. This ensures proper compensation for services and helps improve revenue cycle efficiency by reducing claim denials or delays related to eligibility issues.
Enhancing Prior Authorization: Accelerate the precision of prior authorizations via AI algorithms that evaluate candidate patient introductions with predefined guidelines and clinical standards. This leads to better decisions with less information, streamlining procedural control automation.
Schedule Filling: AI-driven scheduling systems intelligently assess patient demand, physician availability, and resource utilization data to optimize appointment schedules, minimizing no-show rates and optimizing resource utilization.
Optimizing Billing and Coding Accuracy: Automating coding and billing procedures improves accuracy and speeds up reimbursement cycles, reducing errors and denials in claims.
Enhance Claim Processing and Detect Errors: Advanced algorithms automate claim processing and help lower personnel costs by reducing the need for additional RCM staff.
Minimized Denials: AI recognizes trends in claim rejections by investigating historical claims data management, limiting revenue loss due to denials, while enforcing remediation steps for charge capture and documentation optimization.
Investigate the Revenue Cycle: AI algorithms combined with automation scrutinize vast amounts of healthcare data to detect trends regarding insurance claims, reimbursement rates, patient profiles, and physician performance, informing financial decisions.
How Traditional RCM AI Lowers Transaction Costs
Healthcare executives are turning to standard AI for financial and operational reprieve, facing staffing shortages and rising clinician burnout. Currently, many healthcare organizations are operating at a deficit. The average hospital profit margin is now 3 percent, compared to 7 percent in 2019, as reported by Moody's. According to Fitch Ratings, 50% of hospitals are either unprofitable or failing to produce any revenue.

Cost reduction is essential for sustainable operation. By combining traditional AI with automation, labor costs, prior authorization, and claim denials can significantly decrease. McKinsey & Co. estimates that the U.S. healthcare sector could save between $200 to $360 billion annually through intelligent automation and traditional AI on administrative tasks. Research from the Institute for Robotic Process Automation and Artificial Intelligence supports this, indicating that healthcare costs could be slashed by 25–50% when combining traditional AI with automation.

Improving Working Conditions with AI
Beyond cost savings, traditional AI is also enhancing working conditions within the revenue cycle. The Association of American Medical Colleges predicts a physician deficit of up to 124,000 by 2034, citing increasing complexity in documentation and job dissatisfaction among doctors.

AI and automation can alleviate the burdens placed on revenue cycle staff, whose workloads often include extensive tasks like prior authorizations and eligibility verifications. Many RCM workers might consider more manageable jobs outside their field, as indicated by a survey where 49% of respondents were contemplating leaving for other healthcare positions.

Utilizing AI and automation-enabled software can improve job satisfaction and productivity among staff. A survey conducted by Salesforce found that 89% of automation users reported increased job satisfaction, and 84% noted improvements across the company. Ultimately, enhanced job satisfaction can reduce turnover rates, benefiting overall operational costs.

Generative AI Approaches in Healthcare RCM
While we are still in the developmental phase for generative AI in revenue cycles, it is crucial to anticipate future trends to position your organization for cost-effective technology adoption. Technology shifts can drastically alter industries, as seen with Blockbuster and Kodak's failures to adapt.

Promising Uses for Generative AI
Here are some of the most effective uses of generative AI emerging in the healthcare sector:

Creating automated clinical documentation.
Developing real-time patient interaction bots for FAQs and appointment scheduling.
Generating patient discharge instructions tailored to individual needs.
Assisting with data collection and analysis for clinical decision support.
Advancements in Medical Records Management
Development of Software for Note-Reading and Voice-Recognition Dictation: Using deep learning powered by generative AI, software is now able to read physician notes faster and with greater accuracy. However, the implementation of this technology has proven challenging for many physicians, with few health systems achieving moderate success in automating physician notes using OCR and natural language processing (NNLP).
Challenges with Traditional OCR: While traditional AI has provided OCR solutions, these systems often struggle with errors, especially with documents featuring complex layouts or messy handwriting. Generative AI significantly enhances OCR capabilities, allowing for better contextual understanding and improved accuracy in processing partially visible or poorly written text.
Enhanced Note-Taking and Transcription: Generative AI software aids in more precise note-taking and transcription during patient interactions, automatically using that information to support administrative billing events. This alleviates the burden of note-taking from physicians, reducing "pajama time" spent on documentation.
Improving Patient Accessibility
Recent advancements involve the use of generative AI to identify redundant patient records early in their interaction with the system. This extends beyond eligibility assessments to automate and align them more closely with specific payer policies and agreements.

Currently, traditional AI-powered solutions can assist with some prior authorization tasks, but generative AI enhances accuracy in reviewing patient data, medical histories, and insurance information. This allows for quicker determination of whether treatment criteria are satisfied and helps mitigate operational staff burdens by addressing time-consuming exceptions.

Transforming Receivables with AI
Developers are exploring how AI-generated material can streamline accounts receivable in healthcare organizations. By utilizing historical performance metrics and payer policy information, generative AI could automate the creation of unique appeal letters for health insurers. This system acts similarly to ChatGPT for appeals, automating all outreach efforts.

Debating Generative AI's Role in Healthcare
Generative AI is emerging as a significant player in the healthcare sector, and organizations must consider whether they will adapt quickly or take a more cautious approach. A survey by Bain & Company revealed that 75% of healthcare executives believe generative AI is set to reshape the industry, yet only 6% have established a strategy to implement it.

Understanding the Risks of Generative AI
Despite its promise, early adopters highlight important concerns surrounding generative AI:

Accuracy and Robustness: There are apprehensions about generative AI's ability to produce accurate outputs, as errors could jeopardize the revenue cycle and lead to suboptimal patient outcomes.
Ethical and Legal Challenges: AI systems often mirror the biases present in healthcare billing practices, raising significant ethical and legal concerns that need to be addressed.
Patient Data Security: The reliance on patient data for accurate insights makes AI systems vulnerable to hacking, thereby compromising patient privacy and data security.
Contextual Understanding Limitations: Generative AI lacks the contextual understanding that experienced healthcare professionals possess, which can lead to missed insights vital for nuanced clinical decisions.
Integration Challenges: Implementing generative AI requires considerable time, effort, and financial investment, along with potential resistance from staff.
Strategies for Successful AI Implementation
Bain & Company suggests several best practices for healthcare systems looking to harness the potential of AI:

Ensure a strong commitment to technology initiatives guided by a clear end-state vision through comprehensive planning.
Design robust processes for redesigning operational procedures for maximum value capture.
Develop a detailed technology investment strategy, ensuring funding for projects and maintaining quality assurance throughout implementation.
Establish effective feedback loops and holistic outcome-based success metrics to gauge performance and success effectively.
The Future of AI in Revenue Cycle Management
As interest in AI quickens, revenue cycle professionals are excited by headlines touting AI’s transformative potential in healthcare. However, it’s critical for healthcare vendors to distinguish between traditional machine learning applications and the advancements brought forth by generative AI.

Currently, while generative AI is a promising technology, its implementation in clinical settings may still lag behind the hype. Healthcare revenue cycles are inherently complex and require nuanced understanding beyond what a generative AI can provide. Nevertheless, both traditional AI and generative AI offer opportunities for improving performance within healthcare systems.

Regardless of whether you lean toward traditional AI solutions or the emerging capabilities of generative AI, understanding your organization's financial landscape is vital to identify underpayment areas and drive revenue improvements.

For further insights into utilizing AI for enhancing revenue cycle management, be sure to visit Velan Healthcare Solutions. To learn more about the evolving landscape of AI in healthcare, see the latest updates from Bain & Company.