Artificial intelligence solutions can reduce the administrative burden of pre-approval and billing, but organizations report increased transaction volumes and costs, according to a new report.
The Peterson Institute for Health Technology gained insights from a January 2026 workshop attended by senior leaders from organizations ranging from health systems to federal agencies. Leaders discussed how technology and policy can enable AI to reduce administrative costs, accelerate payment cycles, and promote high-value care.
While AI has the potential to help organizations perform faster pre-authorizations at lower costs, the report (PDF) states that there is no existing evidence that AI “leads to lower average cost per claim, given the cost of the AI solution.”
Participants also cited increased system activity, including back-and-forth “bot wars,” as well as limited impact on complex cases and unintended consequences, as potential risks of deploying AI with pre-approval.
The report also notes that while real-time pre-authorization at the point of care is an emerging model, it is not currently scalable and the impact of AI will be limited by policy changes. Participants provided several suggestions for improving current data standardization, including requirements for the integration of standard prior authorization APIs by electronic health record (EHR) vendors and extending requirements for standardized electronic transaction types to additional health plan types.
Examining the role of AI in medical billing, the report found that provider adoption is “increasing billing intensity and healthcare spending,” with AI scribes in particular increasing billing intensity in add-on codes for assessment and management and diagnosis-related group complexity.
“AI is accelerating the growth of more complex claims, which has already put pressure on affordability in recent years,” the report said. “The result is a trajectory that payers and patients cannot continue to absorb.”
AI tools are becoming commonplace in healthcare organizations across the U.S. A March study from Eliciting Insights found that 75% of healthcare systems are using at least one AI platform, up from 59% in 2025. Additionally, 50% of respondents said they use three or more AI applications in their systems.
As adoption increases, systems report significant barriers to implementation efforts, with 74% of respondents citing EHR vendor dependence as a barrier in the Qventus report.
The report said health plans are making “total downcoding” and other reimbursement cuts as claims intensify. However, the impact of such reductions is currently unknown and could “unduly harm” providers who have not yet deployed AI tools.
The report also said that current health plans are “likely not sufficient to address AI-driven health inflation” and called for coordinated policies to address it.
“This argument confirms a core reality: AI in healthcare management processes, as currently implemented, will likely only achieve some of its goals, such as reducing manual effort for organizations to fulfill prior authorization requests and submit claims, while also increasing healthcare costs,” PHTI executives said in the report.
Furthermore, given workflows, data complexity, and incentives, AI “exacerbates the underlying problems,” the report said. Ultimately, researchers say the deployment process needs to be redesigned to reduce administrative waste.
