Agentic Workflows
AI agents that talk to patients, answer questions from the chart, and take care of the busywork — with safety guardrails built in.
About Agentic
Agentic workflows put an AI agent in the loop — one that can understand a request, decide what to do, call tools, and respond on its own. In healthcare that's powerful and risky, so these workflows pair LLM reasoning with hard guardrails.
From a voice agent that calls discharged patients to an assistant that answers questions from the live chart, each flow combines an AI model, the tools it can use (FHIR search, eligibility checks, phone calls), and safety layers like PHI redaction and allowed-resource checks.
What you can do
- LLM-powered agents that reason, use tools, and take action
- Built-in guardrails — PHI redaction, allow-lists, red-flag escalation
- Voice and chat — call patients or answer questions from the chart
- Grounded in real EHR data, not guesses
Requirements
Set up these connections in the Workflow Builder before running any workflow in this category.
EHR Connection
Agents read and write patient data through the EHR's FHIR API or SMART embed context. Backend credentials for data queries; embed credentials for patient-facing launch contexts.
LLM Provider
An AI model provider (AWS Bedrock, Anthropic API, etc.) that powers the agent's reasoning, tool selection, and natural-language responses.
Telephony Provider (voice workflows only)
A voice API such as Twilio required by workflows that make or receive phone calls. Provides the real-time audio stream the AI agent listens and speaks over.
Workflows
2 ready-to-use workflows — open any to see the full flow and download it.
AI Voice Agent for Patient Discharge Calls
An AI voice agent that calls discharged patients, runs a check-in questionnaire in natural conversation, and escalates red flags to a nurse.
AI Clinical Assistant, Powered by Live FHIR
An AI agent that answers plain-English questions from live FHIR data — with allow-list and PHI-redaction guardrails built in.