Top use cases for agentic AI in revenue cycle management
High-value use cases for agentic AI in revenue cycle management are emerging as early adopters realize greater efficiencies, automation and payment accuracy.
Editor's Note: This is the second article in a three-part series on agentic AI in revenue cycle management. Find the first article here and stay tuned for another article on implementation considerations.
Despite being in the early adoption phase, agentic AI offers several high-value use cases in revenue cycle management as technology vendors and providers continue to invest in the latest evolution of artificial intelligence.
Use cases for agentic AI in revenue cycle management primarily focus on enhancing efficiencies, accuracy and automation to improve financial health and ultimately the patient experience. Key areas in which agentic AI is applied at this point include claims management, prior authorization, denial management and patient financial engagement.
Medical professionals are putting their faith in agentic AI solutions. According to a recent Salesforce survey of 500 healthcare professionals, AI agents could cut administrative burden by 30% for doctors, 39% for nurses and 28% for administrative staff.
But revenue cycle technology vendors don't believe it will stop there. Agentic AI is poised to disrupt the revenue cycle. Some experts believe the technology could automate as much as 80% of revenue cycle work.
Patient eligibility and benefits verification
Agentic AI can automate patient eligibility and benefits verification processes, transforming a once very manual, repetitive and labor-intensive task.
Leveraging natural language processing, AI agents can identify and capture the appropriate information from insurance cards, the EHR and other health IT systems and query payer systems for verification.
Using agentic AI for patient eligibility and benefits verification can accelerate the revenue cycle and reduce the likelihood of denials from eligibility errors. Healthcare organizations can also increase the number of verifications, smoothing medical billing and collections later.
Patient eligibility and benefits verification is also a top use case cited by healthcare professionals in the Salesforce survey, in which 70% said they want to use agents in this and other aspects of healthcare.
Prior authorizations
Prior authorizations are one of the biggest challenges in healthcare right now. The notoriously burdensome task of receiving pre-certification for services has been linked to adverse patient events, delayed care and administrative overload. But agentic AI is aiming to overcome all of that.
Agentic AI can autonomously gather clinical documentation and patient data from EHR systems, review payer policies and requirements, complete and submit authorization forms and track requests. The technology can also do this with minimal human intervention.
What's more, vendors are designing AI agents to proactively identify potential issues impacting a prior authorization's approval, whether it be compliance issues or missing documentation.
Similar to AI-based eligibility and verification, using AI agents for prior authorizations can clear the path for cleaner claims.
"We think about some of the pre-service tasks, so everything from eligibility verification, benefit verification and prior authorization, a lot of those we're seeing a lot of success in," said Dan Parsons, co-founder and chief experience officer at Thoughtful AI, a company that delivers AI-driven revenue cycle technology.
Denials management and appeals
Claim denial rates continue to rise as revenue cycles handle an increasing volume of claims. AI is helping to bridge this gap amid staffing shortages. But agentic AI, in particular, has the potential to revolutionize how revenue cycles manage denials.
Due to its autonomous nature, agentic AI solutions can analyze claim denial codes, identify patterns and trends and retrieve the data needed to correct errors. Agentic AI can also prioritize denials based on their impact on revenue, so healthcare organizations get the most bang for their buck.
Vendors also see an opportunity to automate the appeals process with agentic AI.
In perhaps one of the strongest use cases right now, agentic AI is completing claim denial appeal processes from initial denial to appeal submission.
"We believe that the denials appeal process is another really good use case," explained John Landy, chief technology officer at FinThrive. "If you could think about today, humans are essentially handling denials, calling payers to get the details of what happened, packaging together an appeals packet, and resubmitting the claim to payers. To perform that work at scale for large organizations requires a lot of people."
Agentic AI can eliminate that manual work and scale appeals operations for faster and greater denial overturns.
Claims management
Part of what makes claims management so complicated -- yet ripe for AI disruption -- is the act of parsing through payer contracts.
You got a hockey stick curve of use cases because you got the data, and the tools are incredible now.
John Landy, CTO, FinThrive
Agentic AI can automatically analyze payer contracts to learn the rules and requirements for clean claim submission, then gather the information needed to submit the claim. As the AI agent goes through the process, it can also learn what is likely to be paid or denied and adjust its practices to align with those trends.
"Loading contracts from payers is a really good use case," Landy said. "The industry in general has wanted it performed very quickly, and it is highly manual, so there are a lot of error-prone steps involved in it."
Using agentic AI in claims management can reduce approval times by streamlining and automating claims processing and submission. More accurate pre-bill scrubbing will also result in middle and back-end efficiencies.
Patient financial communications
Consumers encounter AI agents for customer service across many industries. Healthcare can tap into this solution for better patient financial communications. Agentic AI can streamline interactions with healthcare consumers and patients while still delivering a personalized experience.
AI agents can handle routine billing inquiries, which are increasingly being handled online versus by phone or mail. The agents can answer frequently asked questions about charges and payment options, then process payments. The agents can even explain billing statements to patients and deliver personalized billing information, including deductible status and balances owed.
Additionally, agentic AI-powered agents can provide multilingual support.
Vendors are already seeing success with AI agents in their patient contact centers, with Ensemble's Ivy reporting higher one-touch resolution rates. This means more patients have their questions answered upon first contact.
These agents are also typically designed to escalate communications when an AI agent is no longer fulfilling the needs of the user. So, while agentic AI can streamline patient call center workflows, it can also recognize when a human is needed for resolution.
Beyond use cases
While several use cases have emerged during this early adoption phase, this seems to be the tip of the iceberg for agentic AI in the revenue cycle.
"You got a hockey stick curve of use cases because you got the data, and the tools are incredible now," Landy explained. "So, we can scale applications and go from ideation to delivery really quickly. In the past, it used to be a science project, testing models for a year, and sometimes they would work, sometimes they wouldn't. Now, we feel good about not only going after good use cases but getting those to the market much faster than we used to."
However, according to Ivy, agentic AI isn't necessarily just about finding the use cases.
"I hate the word use case because the goal is not to develop a case that works; it's to develop a goal that can be enterprise-deployed," Ivy explained.
The orchestration component of agentic AI lends itself to more widespread application across an organization. When AI agents can talk to other AI agents to not only communicate but also learn from each other, then revenue cycles can move faster with fewer errors trickling from one part to the other.
Jacqueline LaPointe is a graduate of Brandeis University and King's College London. She has been writing about healthcare finance and revenue cycle management since 2016.
