Healthcare AI is leaving its experimental phase behind — most organizations are already using it, and 43% are piloting or testing agentic AI.
There are numerous success stories where healthcare organizations deployed AI in pilots or applied AI in a single department. These successes — which may have involved ambient documentation or use of AI in the revenue cycle — have demonstrated AI’s potential to improve clinical outcomes, decrease costs and deliver positive ROI.
However, these successes have often been limited to small, isolated silos. Meanwhile, only 4% of healthcare organizations report successfully scaling AI across the entire enterprise.
These struggles in generating enterprise-wide value from AI are similar to challenges experienced in other industries. According to Gartner, at least half of GenAI projects are abandoned after proof of concept, and an MIT report found that 95% of GenAI projects fail.
Healthcare leaders — especially those at organizations with successful AI pilots — are wondering what it takes to successfully scale AI across the enterprise.
A framework to operationalize AI in healthcare
UiPath, with a long history in the healthcare automation space, is focused on helping organizations automate processes. Based on its experience, UiPath has developed a framework to guide healthcare organizations in operationalizing AI across the entire enterprise.
First, before automating or applying technologies like agentic AI, it is necessary to identify every step in the organization’s key processes, end to end.
“Some of the most complex processes in healthcare might have 10, 20, 50 or even 100 different individual steps,” said Chris Jordan, director of AI solutions in healthcare and life sciences at UiPath. “Those steps typically span different stakeholders, in different departments.”
The main steps in most processes commonly encompass four key areas: 1) ingestion of data; 2) research and analysis, which is the most important cognitive step; 3) data entry, as data is entered into a system, such as an EHR and 4) resolution and close.
The framework developed by UiPath then uses different technologies to automate these steps and end-to-end processes.
These technologies include intelligent documentation processing (IDP) tools that enable pulling data from one document, such as a fax, analyzing it and putting it into the system of record; agentic AI, which uses agents to perform specific tasks — including cognitive tasks that previously required a human — and robotic process automation (RPA) for simple, deterministic, data-entry tasks.
The final “technology” is the human in the loop. “Despite the use of AI and agents and other exciting technologies, we still need humans to be part of the process,” Mr. Jordan observed. “The goal is to reduce the workload on them so they can review the output of the technology and decide whether it is correct or not.”
The most important factor for operationalizing AI: Orchestration
Mapping the steps and implementing specific technologies is important but not enough. Successfully operationalizing AI across an entire enterprise requires an additional technology — agentic orchestration — that overlays and directs all the other technologies.
The many tools that enterprises are using include IDPs, bots and multiple agents, including agents that can perform increasingly complex reasoning. Each of these tools performs specific tasks. These tools need to be orchestrated so they perform these tasks in exactly the right sequence, at exactly the right time in the workflow.
“Similar to the idea of right care, right place, right time — which has always been a core principle in healthcare — we need agentic orchestration to get the right tools intervening in the process at the right points in time,” said Peter Reischer, senior healthcare solution engineer at UiPath.
UiPath has developed an orchestration platform for healthcare that is vendor agnostic. It can work with API automations, FHIR integrations, EDI integrations from prominent EHRs and other technologies within a health system’s tech stack.
“The vendor-agnostic approach allows a health system to put the best tool in the right place in its workflow,” Mr. Reischer said. UiPath’s orchestration platform integrates with these tools and enables automation of complex end-to-end processes.
An orchestration platform ensures getting the right data into AI, getting the right data out of AI and providing improved visibility of an entire process.
“Without orchestration, you’ve really just got this assortment of various point solutions that don’t all work together and you don’t really have an understanding of how you’re performing,” Mr. Jordan noted.
Realizing value from AI at scale
The goal is to go beyond successful but fragmented pilots to implementing AI across the entire enterprise.
The combination of understanding the organizations key end-to-end processes, applying UiPath’s framework, adopting an orchestration platform and keeping a human in the loop for key decisions makes it possible to operationalize AI enterprise-wide.
Once AI is operationalized for key processes and workflows, they’ll be able to drive real value, reduce the administrative burden and improve clinical and financial performance.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.