How AI is improving revenue cycle management

Where AI exceeds expectations

Greater automated revenue capture and coding produced the most benefits, according to survey respondents. Eighty-four percent of respondents said AI for revenue cycle management met or exceeded expectations for more automated revenue capture and coding, with more saying it exceeded expectations (53% versus 31% of met expectations and 16% of below expectations).

Finance and IT leaders also found that revenue cycle AI led to increased automation for claims management. Half of the respondents said AI exceeded expectations for more claims management automation, while 32% said it met expectations. Just 18% said it did not meet expectations.

Another area with a high level of results was analytics and reporting. A total of 89% of the respondents felt AI met or exceeded expectations -- 43% and 46%, respectively -- with improving analytics and reporting through machine learning or predictive analytics. The least number of respondents also said AI did not meet expectations in their area, with just 11% of the respondents.

Most respondents also felt AI exceeded expectations (versus met expectations or did not exceed expectations) for increasing the accuracy of transactions or producing fewer errors and increasing the speed of patient eligibility verification.

However, revenue cycle AI has some work to deliver on the expectations providers have for improved denials management and increased security within revenue cycle management workflows. Respondents felt AI largely met expectations in these areas, but more respondents felt the solutions were below expectations compared to other metrics (27% and 23%, respectively).

Beyond hype

The survey results suggest that AI for revenue cycle management is no longer just hype. AI capabilities are improving revenue cycle workflows and streamlining the medical billing process. As providers realize these gains, confidence in the technology is also growing.

In the survey, 60% of respondents said they have more confidence today in AI that can support the revenue cycle effectively compared to when they first adopted AI capabilities.

Trust in AI for revenue cycle management is still a challenge. About 39% of the respondents said that trust in AI output is an obstacle to deciding on AI investments for revenue cycle management. Although nearly 60% said their mistrust has decreased significantly.

Healthcare finance and IT leaders can also rely more on AI output. The survey found that few respondents felt AI for healthcare payments, specifically, were less accurate than when a human did them (8%), while another 8% said it was too soon to tell. Meanwhile, most respondents (41%) felt the AI payments were slightly more accurate, compared to 19% who said they were significantly more accurate and 25% who said it was the same level of accuracy.

With higher levels of confidence, most providers plan to increase their investments in AI capabilities. Fifteen percent of respondents plan to invest significantly more in revenue cycle AI with an expected increase of 10% or more over the next one to two years. Another 44% said they plan to invest 1% to 9% more.

New AI capabilities that providers have set their sights on include enhanced payer engagement, particularly through generative and agentic AI solutions. The metrics providers expect AI to improve include claim rejection rates, net collection rates and cost to collect.

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.