Artificial intelligence in healthcare is rapidly transitioning from passive data analysis to active task automation. Nowhere was this evolution more visible than at HIMSS25, where discussions centred on agentic AI – autonomous agents capable of managing complex workflows with minimal human intervention. As these technologies move beyond ambient listening to orchestrating tasks across electronic health records (EHRs), healthcare organisations face the dual challenge of seizing operational efficiencies while safeguarding clinical decision-making. The road ahead demands both innovation and caution to strike the right balance between automation and human oversight.
Operational Gains and Clinical Limits
The adoption of agentic AI is being driven by a compelling need to reduce the administrative burden on clinicians. These AI agents go beyond basic generative tools by acting on data and initiating follow-up actions, effectively serving as digital collaborators. Vendors like Epic and InterSystems are leveraging this capability to handle pre-visit preparations, patient communications and post-surgical follow-ups. For instance, Epic demonstrated how a patient recovering from wrist surgery could interact with an agent that assesses healing progress via voice and image inputs, referencing population data to suggest whether a follow-up visit is necessary.
Must Read: Harnessing Collaboration and AI for Enhanced Patient Safety
While these examples illustrate a potential paradigm shift in efficiency, they also underscore the importance of human verification. The complexity of clinical care – filled with nuance, judgment and unpredictability – remains a boundary that autonomous systems have yet to safely cross. There is a collective consensus among technologists and healthcare leaders that while agentic AI can accelerate and augment clinical workflows, it should not replace physician decision-making. The potential consequences of over-automation – such as biased decision-making or patient safety risks – necessitate a deliberate and measured implementation strategy.
Governance, Caution and Gradual Integration
The rapid pace of AI development has outstripped existing regulatory frameworks, which were built for earlier, more limited forms of machine learning. HIMSS25 spotlighted this governance gap, with experts highlighting the urgent need for new policies to manage the rise of autonomous systems. While some legislative proposals have controversially explored recognising AI as eligible for medical prescribing under FDA guidelines, many technology leaders argue the field is not yet ready for such a leap.
Healthcare organisations are, therefore, moving cautiously, with many vendors confining agentic AI applications to administrative functions. InterSystems, for example, introduced IntelliCare, an AI-powered EHR that prepopulates billing codes and generates patient history summaries but avoids autonomous clinical decision-making. Similarly, eClinicalWorks has prioritised RCM and document processing, deploying AI to handle tasks like extracting data from PDFs and escalating complex queries to human agents. These approaches reflect a broader industry commitment to transparency, trust-building and human oversight – all essential for responsible AI deployment.
Revenue Cycle Management as a Launchpad
The most immediate and tangible value of agentic AI is unfolding in the realm of revenue cycle management. As healthcare organisations face increasing pressure to streamline financial operations, AI agents are being used to automate claim processing, coding verification, prior authorisations and other repetitive administrative tasks. Companies like athenahealth and eClinicalWorks are applying agentic workflows to reduce friction between payers and providers, improve task navigation and achieve significant productivity gains.
This trend is not only easing the workload for human staff but also enabling organisations to repurpose teams toward more strategic functions. By delegating routine back-office processes to AI agents, some RCM departments are witnessing unprecedented efficiency improvements. These developments signal a shift in how healthcare organisations view administrative automation – not merely as support for clinical care but as a central component of their operational strategy. While the clinical use of agentic AI remains limited, the success of its implementation in RCM may pave the way for broader adoption across healthcare systems.
Agentic AI represents a powerful new tool in the digital transformation of healthcare, capable of coordinating and executing tasks across EHRs and beyond. The innovations showcased at HIMSS25 demonstrated clear benefits in improving efficiency, particularly in administrative and revenue-focused functions. However, the cautious tone adopted by many industry leaders reinforces a crucial point: in clinical care, the stakes are too high for unchecked automation. As agentic AI continues to evolve, the path forward must be guided by strong governance, human oversight and a commitment to patient safety. The healthcare industry stands at the threshold of a new era – one that demands both technological ambition and ethical responsibility.
Source: Healthcare IT News
Image Credit: Freepik
