Healthcare leaders expect AI maturity this year, especially in RCM
A new report shows low AI maturity despite widespread adoption in the revenue cycle. Still, leaders expect to mature rapidly over the next year or two as they confront data constraints.
This year, the healthcare system is at a major crossroads in its AI journey. The application of the technology will either solve fragmentation, particularly in the revenue cycle, or exacerbate this major challenge for healthcare providers.
That is the takeaway from Innovaccer's latest report, "State of Revenue Lifecycle in Healthcare," based on a survey of 150 U.S. healthcare leaders. These leaders revealed a turning point for payers and providers using AI, as they move from implementations and experimentations to a more cohesive, scalable AI strategy.
But right now, about seven in 10 leaders identify their organizations as in the early or mid-stage of AI maturity, despite most already using AI across multiple workflows, particularly in clinical documentation and coding support, patient access and scheduling, operational workflow automations and revenue cycle activities.
Meanwhile, just 8% identified their organizations at the top of the scale with enterprise adoption. Only 22% more said their organizations were in an advanced stage or AI-native.
However, the survey suggested that healthcare leaders are ready to evolve their AI capabilities, seeking to level up over the next two years. This could bring about significant change to the revenue cycle if organizations can overcome the primary barriers, Abhinav Shashank, cofounder and CEO of Innovaccer, told RevCycle Management.
The state of AI in healthcare right now
Financial pressures on healthcare organizations will push leaders to advance their AI capabilities and do so quickly, Shashank explained.
"These [health] systems are under so much financial stress, thinking about administrative processes, and simplifying that is just basically going to be an incredibly important theme," he said.
Leaders are ready to take on the challenge; most organizations expect to move up at least one level in AI maturity over the next two years, according to the survey. This means maturing from tool experimentation and functional adoption of AI across multiple use cases to enterprise integration and perhaps beyond.
[T]here is going to be a massive drop in the cost structure, not just in how you process revenue, but also a bunch of administrative simplification as this goes on.
Abhinav Shashank, cofounder& CEO, Innovaccer
For example, 52% of survey respondents said their organizations use AI for workflow, while another 46% use it for documentation support.
Other common domains for AI applications right now include scheduling and patient access (41%) and revenue cycle automation (38%). Examples of how AI is being applied in these domains include AI-assisted coding, automated claim edits before submission, denial-prediction models that flag high-risk claims and propensity-to-pay scoring.
These AI capabilities are also generating ROI for healthcare organizations, turning the technology into what Shashank calls the "financial backbone of health systems."
Survey respondents reported that AI has resulted in up to a 40% reduction in documentation time, 25% fewer denials and 45% faster scheduling. Respondents also said AI capabilities have lowered call center volume by 30%.
But healthcare organizations can achieve significantly greater returns from their AI investments in administration and revenue cycle management, Shashank stated.
"People are doing that in point pieces. First, trying to figure out how to scribe better, then trying to figure out how to code better, et cetera. But our general view from this [report] is that all of this is going to come together, and I think pretty quickly over the next year or so," he explained.
"Once it does, there is going to be a massive drop in the cost structure, not just in how you process revenue, but also a bunch of administrative simplification as this goes on."
However, healthcare organizations will need to address a major barrier to advancing to the next level of AI maturity and eventually reaching the ultimate goal: autonomous healthcare administration.
Will AI solve fragmentation or supercharge it?
Data fragmentation is the top challenge of becoming more AI-mature . Approximately 62% of leaders in the survey cited fragmented data systems as the primary barrier to scaling AI capabilities within their organizations, surpassing staffing shortages, model trust, budget constraints and transparency concerns.
Health IT infrastructure must be a top priority when implementing and scaling AI to achieve enterprise efficiency gains, according to the survey. Yet, clinical, financial and operations data generally live in different systems, formats and ownership structures.
Healthcare leaders must make sure their ongoing AI implementations do not exacerbate this data fragmentation problem.
[I]f we thought EHR interoperability was a hard problem to solve, imagine getting 200 of these agents interoperable with each other.
Abhinav Shashank, co-founder& CEO, Innovaccer
"Due to the AI wave that we are in now, we are creating [AI] agents for everything," Shashank stated. "So, someone is going to vendor A to do scribing, then vendor B to do coding, and vendor C to do prior authorizations. Suddenly, you have 20 things in your tech stack, which will not really communicate with each other."
"And if we thought EHR interoperability was a hard problem to solve, imagine getting 200 of these agents interoperable with each other," he continued.
This fragmentation of systems and data prevents enterprise-level AI maturity and systemic automation that can scale administrative simplification and financial results.
The healthcare industry is still riding this wave of AI adoption, but it will come crashing soon, Shashank suggested.
"Once this cycle evolves, we get to a point where there is an emergence of platforms that can really get all of this to work together and have a common data infrastructure, common systems in place," he said.
The report describes this as a platform-first approach to AI.
A platform-first approach supports AI unification. When AI capabilities can share data and identity layers, organizations can achieve compounding ROI, consistent behavior across care settings, composable workflows, common context across all copilots, standardized safety and enterprise-level governance, the report explained.
Shashank likened this approach to how Google was able to bring disparate websites together on the Internet for people to access more.
"I think there is going to be a similar platform emergence on the revenue cycle side, and the administrative side of healthcare," he said. "There will be companies that effectively get all these agents together, and once you have all of these point solutions connected, running on a common fabric and chassis, you are really going to see these things start to work like magic and really drive deflationary effects on the administrative cost structure for health systems."
2026 will be the crossroads for AI maturity
Healthcare organizations are at a crossroads now. Do they keep adding point solutions to solve specific challenges within their organizations and continue to see some ROI? Or do they apply a platform-first approach to get to systemic automation and possibly an autonomous revenue cycle?
AI maturity starts with systemic automation whereas the first levels are tool experimentation, functional adoption and enterprise integration. By systemic automation, workflows become predictive and increasingly prescriptive. Identity and security are also robust alongside AI risk management structures. Platform adoption is there.
The final level is autonomous healthcare administration. In this case, AI managed high-volume administrative tasks with human oversight, including routing, coding, scheduling, triage and parts of the revenue cycle. These operate in a closed-loop fashion, and staff focus on edge cases and relationships.
These higher levels of AI maturity are not defined by acquiring more tools but "building a unified data and identity layer, adopting a platform strategy, and instituting AI governance."
Healthcare leaders are putting in the work here. About 45% of leaders said they have established AI governance/ethics structures, and 52% have expanded implementation across departments.
In the next 12 to 24 months, about 44% of leaders also said they are investing in unified data platforms, while 45% are strengthening governance. Still, 52% said they are focusing on workflow and revenue cycle automation, and 46% are prioritizing predictive analytics this year.
Healthcare leaders generally do not expect to have autonomous healthcare administration even in the next three years, and the report puts that closer to 2030 -- if healthcare leaders strive for a unified AI strategy.
"For you to succeed in that in the long-term, you will need a strong data infrastructure to really exist," Shashank said. "And the organizations that are spending time, effort and energy in setting up that infrastructure today are going to be the winners over the next two to three years."
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.
