Developing an AI Strategy within Revenue Cycle Management

There is an ongoing war between Healthcare Providers and Payers. A majority of insurance companies are using AI to process and deny claims. Generally speaking, providers are using AI at a very limited capacity, if at all. And frankly, this puts providers and thereby their patients at a serious disadvantage. The issues are well-known: timely filing, prior/pre-authorization submission/status, medical coding discrepancies, and widespread work force challenges. Simply adding more bodies or automating the problem isn’t enough any longer. There are too many changes and variances for a provider to keep up with in Medicaid, Medicare, and commercial insurance plans.  

Unlike a payor who has ONE plan to manage, a provider has MANY plans to react to and make adjustments for.  

So, what can be done? The answer is AI Automation. Using this technology, one can automate the business process needed and capture data throughout the process. This data can then be leveraged to provide internal intelligence to predict, act and ensure that an authorization and/or claim is successful.  

There are several steps to this approach that will ensure that not only is the process automated, but that it is done so with intelligence to gain greater success, less manual intervention, and timely filing over time. This also makes the solution more nimble and agile to adapt to the changes and variances that are inevitable with payors.

CampTek’s solution is composed of several parts:

     I.  Automation and AI First. Using AI, we can take a historical view of the 835 files from top payors to identify high-level trends around denials by payor, CPT code, diagnosis, physician and patient demographics.

Using this data, we build Prior Authorization Automations that perform both the request and status parts of the workflow. Our data team begins to pull the data from the automations to feed the AI Model. These automations are bespoke and will work any EMR, Payor Portal, Availity, Waystar, or other clearing houses.

See examples of the Prior Auth Request and Status Bots running in action

     II.    Intelligence Building. Once the automations are running in production, we begin gather data and intelligence around the authorizations that are getting statused. This data will be fed to the data model continuously. In addition, any changes in the payors plan that are communicated via their website, pdfs, or other forms are also fed to the data model to make the AI more intelligent.

     III.   Intelligence Consumption. This approach, in time, will provide incredible value in that it can react and adapt to payor changes, but also ensure that the prior authorization submitted will be completed and have a higher chance of acceptance. The outcome will be capturing lost top line revenue and providing specific in-house intelligence for prior authorization using predictive, generative and analytic AI intelligence. This can be consumed in a platform independent way (agentic AI Bots, data analytics programs, ChatBots, Epic or other applications that have API integration).

This is a feed and use the “brain” approach all while giving an operations team the tools it needs to be efficient, save on costs and capture the revenue that would otherwise be written off. Payers are increasingly using AI, so this is a crucial strategy for providers to have this on their roadmap, so as not to be left behind.

Overview of the Full Prior Authorization AI Solution