AI and account resolution: Achieving real results

Ken Ruiz, Ryan Hartman, Caroline Lalla
| 11/6/2023
Kodiak Solutions

Artificial intelligence and machine learning are revolutionizing account resolution to help organizations achieve the most value – even with limited resources.

It’s a common business challenge for healthcare leaders: How can they increase efficiency and standardization in account resolution while minimizing labor costs and maximizing cash to net revenue?

Luckily for leaders today, data science offers a promising solution to this challenge by removing some of the most manual, repetitive account resolution tasks and replacing them with standardization – and a scalable approach. After hearing about them for so long, more and more healthcare organizations are finally realizing the benefits of artificial intelligence (AI) tools for account resolution. But as with many other forms of technology, they’re using the tools with varying degrees of success. So much more potential is yet to be unleashed.

As healthcare organizations move away from manual debit and credit resolution and toward automation, the following are just some examples of how AI and machine learning are revolutionizing account resolution to help organizations achieve the most value with the limited resources they have.

Use the power of AI to resolve patient accounts faster, cheaper, and more accurately than doing it manually.

Automation’s promise in account resolution: 3 examples

When determining areas in which to apply automation, healthcare leaders should consider those that are high touch and inefficient, involve a lot of manual effort, lack established standardized processes, and provide financial benefit for the organization. Credits, debits, and denials are three examples of such revenue cycle areas.

Credit balances

For most organizations, credit balances require significant manual effort to resolve; on average, it can take a staff member 15 minutes to work a single credit balance account.1 Organizations that are short on full-time personnel dedicated to resolving credit balances face even greater challenges staying on top of accounts.

Some organizations might not think credit balances are a high priority because they don’t generate cash. If organizations do not make credit balances a priority, however, the number of accounts to be resolved can grow to the point where it is unmanageable and takes far too many employees and resources to clean up the credit balance population.

Where’s the automation opportunity?

Automation can help healthcare organizations manage credit balances more quickly and efficiently, which can positively affect their financial bottom line and improve patient satisfaction. Automation is particularly efficient at sorting through larger credit balance populations and can segment credit balances more effectively. Machine learning can help revenue cycle teams identify transactional patterns in credit balance accounts and can use historical patient accounting system (PAS) data to predict resolutions and provide insights on future accounts that might have similar patterns.

In addition, automation can be used to identify various types of credit balances, including those that might generate income for the organization. For example, many states have statutes that allow organizations to take commercial overpayments that have aged past a certain time limit as income rather than having to refund those dollars. Organizations could use machine learning models to detect the aged accounts.

False credits are another area that can benefit from automation. When an organization has credit balances that are not caused by an overpayment of dollars, it can reverse the adjustment transaction and remove the credit balance from the organization’s liabilities without having to send any dollars out the door. Examples of these credit balances include duplicate contractuals or administrative and denial adjustments.

Using automation in account resolution also presents opportunities to increase patient satisfaction. For example, automation allows revenue cycle teams to identify overpayments more quickly so they can be refunded to patients in a timelier manner. Automation also can be used to identify instances in the PAS of patient overpayment and a corresponding patient underpayment. In this case, an organization could resolve a credit and debit balance at the same time, transferring dollars in house without having to bother the patient with a refund and a bill.

Debits

A good practice for working debit accounts today is for revenue cycle teams to focus on the accounts with the greatest expected value. As with credit accounts, doing so helps use limited resources as effectively as possible to generate the most financial benefit for the organization.

Where’s the automation opportunity?

Machine learning can pull together myriad data elements – more than a human can analyze effectively – to identify debit accounts with cash opportunities. Revenue cycle teams can use machine learning to identify what the organization historically has been paid on various types of accounts so they can adjust their efforts away from low-value accounts and put their energy toward higher-value ones.

For example, teams can use machine learning to review daily transactions on patient accounts and calculate statistics such as amount to pay and time to resolve – useful data points when analyzing accounts’ value. Machine learning models can adjust quickly to new information, such as changes to payor contracts, to make predictions about where value in account resolution, including debit accounts, might reside. Overall, automation improves revenue cycle teams’ workflows by isolating lower-value accounts and removing them from the team’s work queue.

Denials

As with credit and debit accounts, machine learning models can help steer revenue cycle teams toward denials accounts that are highest value and therefore worth pursuing.

Where’s the automation opportunity?

Machine learning models can make useful predictions on accounts that have been denied, guiding the team toward those accounts that have additional payments worth pursuing. These models can answer helpful questions such as:

  • Who will pay (the patient or payor)?
  • How much will the patient or payor pay?
  • When will the patient or payor pay?

Machine learning models can then steer high-value accounts right into the team’s work queue.

Achieving the most value

Organizations today have so much valuable data in their patient accounting systems – from remittances to transactions and charges and beyond. Automation offers much promise – on both the front and back ends of the revenue cycle – to use this data.

To help identify where best to focus their resources and energy and achieve the most value in account resolution, healthcare leaders should consider the data they have available to them and the human resources they have who can generate beneficial analytics. For even more help aligning their people, data, and other resources for organizational success, organizations can consider reaching out to data science specialists who can take a holistic view of account resolution and look for improvement opportunities.

1 Based on data from Crowe client engagements, 2023.

Contact us

Ken Ruiz
Ken Ruiz
Chief Revenue Officer, Kodiak Solutions
Ryan Hartman
Ryan Hartman
Director, Revenue Cycle, Kodiak Solutions 
Caroline Lalla
Caroline Lalla