By Noel Felipe, SVP and Revenue Cycle Practice Leader, Firstsource
LinkedIn: Noel Felipe
LinkedIn: Firstsource
With high labor costs, low margins and claims denial rates averaging almost 15%, according to the American Hospital Association, providers need new approaches to effective revenue cycle management. Automation and AI tools can increase productivity, reduce the cost to collect and accelerate revenues. The keys to success today are agentic workflows and generative AI.
Advancing AI Practices Streamline Entire Revenue Cycle
Agentic workflows are a major development in AI. In an agentic workflow, multiple AI agents, each trained in a specific function, automatically work together across systems and platforms, sharing data, communicating and adapting to new situations. Agentic workflows will redefine the depth to which providers can automate the revenue cycle, from preregistration activities to collections.
Generative AI’s role continues to expand, with its ability to read and summarize clinical notes, call histories, payer contracts, evidence-based guidelines and more. These capabilities mean staff spend less time looking up data and carrying out repetitive tasks and can focus on more complex, even higher value work.
Finally, AI co-pilots work alongside humans, shortening tasks and drawing on human expertise. A co-pilot might prompt a next-best action or offer a claim or email for review.
Here are some ways providers use these tools across their revenue cycles.
- Eligibility verification. Providers can improve their revenue cycle at the outset by offering patients convenient self-scheduling and preregistration options with built-in AI agents verifying benefits coverage. AI co-pilots can guide registration specialists to help patients understand their estimated co-payment and charge amounts and even set up payment options.
- Prior authorization. An agentic workflow can initiate a prior authorization request; check the request against payer guidelines; gather the necessary supporting material; submit the request; and automatically follow up on it. AI agents can even launch an appeals process for denied requests. A prior auth agentic workflow can easily manage dozens of prior authorization requests in a fraction of the time clinical staff need to search for and submit required documents and data.
- Coding. AI agents can transcribe, read and identify diagnostic and procedure codes in physician notes and route these to coding software. A large medical specialty practice improved collections by 80% through autonomous AI coding that identified more codes in clinical notes and applied them accurately.
- Clinical documentation integrity. Gen-AI tools can scan records and identify gaps and potential inaccuracies, prompting action so that clinical documentation is more comprehensive, timely and accurate.
- Claims submission. AI agents can help ensure all charges are captured and increase first-pass throughput. An AI agent trained in payer claim edits can spot a potential error in a claim and query a second AI agent to check a code against the patient’s record while another AI agent checks the payer contract. If the situation is simple, the agents can work together to rectify the error and submit the claim.
- Denials management. Al tools can check on claims status and take appropriate actions, such as generating appeal letters. Tools can also predict a payer’s propensity to pay so providers may effectively prioritize rework queues. AI agents can also uncover root causes of payer denials – which may be payer error. Such evidence can help remediate these issues.
- Patient billing, follow up and collection. Automation tools, such as software bots, can streamline payment reconciliation and cash posting. Agentic workflows can accelerate payments, credits, refunds, etc. For example, a propensity to pay analysis can enable staff to prioritize their follow-up efforts on the accounts most likely to pay.
Gen AI and sentiment analysis capabilities built into AI co-pilots can guide customer service representatives through conversations with patients, helping locate relevant data and call history within seconds and prompting effective next best actions. Gen AI tools can develop personalized training and review 100% of calls and make recommendations for how to improve interactions with patients.
Assembling AI Solutions
AI agents, gen AI, AI co-pilots and other tools build on each other’s capabilities. A front-end function enhanced with AI can help ensure complementary mid and late cycle functions operate more efficiently. That makes it important for providers to have a strategy for where to apply AI in their revenue cycle. Process mining and digital twins can help here.
Process mining enables providers to understand precisely where bottlenecks and errors occur in their revenue cycles. Then providers can simulate the impact of different solutions with a digital twin, a virtual representation of their revenue cycle and its processes. Together, these tools can help providers target specific functions and evaluate the effectiveness of near and longer-term AI implementations on them. This exercise shows providers where to implement AI and automation that will net savings and performance improvements now and will further streamline the functions as the technology continues to advance.