AI systems with autonomous decision-making capabilities serving as the foundation for seamless data exchange and collaboration among different platforms and services.
GE Healthcare, a leading name in the healthcare industry, has unveiled a new Agentic AI designed to revolutionize diagnostic imaging workflows. Simon Rost, the Chief Marketing Officer for the Enterprise Imaging & AI Portfolio at GE Healthcare, is spearheading this innovative development.
The Agentic AI is equipped with advanced features that enable it to maintain context across multiple systems and follow institution-authored policies. This AI agent is not just a tool; it's a digital assistant that can extract data from faxed or portal-submitted requisitions and create orders directly through the Electronic Health Record (EHR) and Radiology Information System (RIS) User Interface (UI).
One of the key aspects of the Agentic AI is its use of named, scoped identities instead of shared logins for agents. This ensures a higher level of security and compliance. The AI can navigate scheduling applications like staff do, booking appointments and updating orders without needing a bespoke scheduling API.
The Agentic AI also employs policies as code for orderable mappings, scheduling rules, and prep criteria. This ensures a standardized and efficient approach to healthcare imaging.
Least-privilege access and role-based access control (RBAC) are used for agent behavior, further enhancing the security and reliability of the AI. Every action of the Agentic AI generates a complete audit trail, ensuring comprehensive auditability.
The AI is designed to move manual touches per workflow, time-to-schedule, no-show rate, repeat scans due to missing priors, and percentage of follow-ups closed on time in the right direction. It does this without introducing new failure modes.
Human-in-the-loop is implemented for ambiguity, with confidence thresholds routing edge cases to staff. The Agentic AI can monitor finalized reports for follow-up recommendations, reconcile them with existing orders, and create the downstream order through the UI.
The Agentic AI can deliver personalized reminders across SMS/email/IVR and write back confirmations or reschedules by following the clinic's UI flow. It can retrieve priors from specialty PACS that use nonstandard exports, converting or exporting using vendor tools.
DICOM/HL7/FHIR remain the backbone, but the Agentic AI offers a third option: Preserve standards wherever they exist and execute staff workflows wherever they don't. This hybrid approach ensures seamless integration with existing systems while automating manual tasks where possible.
Environment isolation and change detection are employed for validating in staging and pausing gracefully if uncertainty spikes. High-cost, well-understood manual workflows are suggested as the first area for modernization, including requisition intake to order creation, priors retrieval and normalization, reminder and prep orchestration, policy-aware self-scheduling for modalities, and follow-up closure and escalation.
The payoff includes fewer manual handoffs, faster cycle times, better scanner utilization, and fewer moments where care stalls because two systems weren't designed to talk. The Agentic AI absorbs variability, turning edge cases into governed workflows that evolve at the speed of operations. It combines language understanding, computer vision, tool use, and policy memory to adapt as interfaces shift.
In essence, the Agentic AI is a game-changer in the healthcare industry, promising to streamline workflows, improve efficiency, and enhance patient care by reducing errors and delays. With GE Healthcare at the helm, the future of healthcare imaging looks promising indeed.
