Agents on FHIR: The Convergence of AI and Interoperability

Details

This content compiles records from a series of 'Agents on FHIR' events held throughout 2026. The central theme is the implementation and application of autonomous agents (AI Agents) using medical data based on the FHIR standard. Specific demonstration examples include Coforix's Clinical Data Transformer and Smart Problem List, and Tokempic, which compresses patient bundles for LLM prompts. Sam Schifman presented agentic cohort analysis over SQL on FHIR and OMOP, and a method to convert clinical guidelines into runnable BPMN care plans. Furthermore, Jens Kristian Villadsen introduced C3PO, an agent that reverse-engineers national health portals into FHIR facades. Aleksandr Kislitsyn argued that FHIR acts as the 'guardrail' preventing AI agents from drifting. The content also covered building FHIR-to-OMOP pipelines and developing open-source audit tools. These examples suggest that combining data standardization (FHIR) with advanced AI techniques (LLM/Agent) enables autonomous clinical support and process automation, going beyond mere data access. It highlights practical implementations such as converting HL7v2 to FHIR and building open-source audit tools.