Building Autonomous Agentic Workflows with the Patient Journey Intelligence Platform

How to build autonomous agentic workflows in healthcare, without exposing sensitive clinical data outside your private environment. Veysel Kocaman shows how Patient Journey Intelligence connects governed clinical data, AI agents, and MCP-based orchestration for real-world healthcare workflows. In this keynote, Veysel Kocaman (CTO, John Snow Labs) demonstrates how healthcare teams can build trusted agentic workflows on top of the Patient Journey Intelligence platform using tools such as Claude Code, Cursor, and Codex. The session shows how multimodal clinical data is ingested, normalised, de-identified, and mapped into a unified patient data layer, then made available through governed agents and MCP endpoints for downstream clinical and research workflows. ⏱️ Key moments: 00:00 What it takes to build agentic workflows in healthcare 02:11 Data ingestion and unifying multimodal clinical data 05:35 Data creation, terminology, and de-identification 08:05 Agents and tools: PJI Assistant, guidelines, and cohorts 12:00 MCP integration: connecting to Claude Code, Cursor, and Codex 14:04 Governance and protecting PHI in practice 16:30 Real-world use cases: trials, care gaps, registries 19:00 Why PJI enables production healthcare AI systems 📌 Applied Healthcare AI Summit 2026 — what actually works in real-world healthcare AI, from pilots to production systems. #HealthcareAI #AgenticAI #AIinHealthcare #ClinicalAI #AIAgents