1 Patient Communications AI Agent, Live in Production.
1 live agent automates patient communications within healthcare. It runs when a message arrives. It publishes the inputs it needs, the steps it works through and what it hands back.
-
Patient Message Triage
Triage an inbound patient message against the practice's own policy, look up the record, and draft either an administrative reply or a clinical escalation. Every reply waits for a human.
Triaging against the practice's own policy, and never replying alone
A practice inbox contains two categories of message that look alike on arrival. Most are administrative — appointment changes, repeat prescription requests, questions about a letter — and answering them is routine. A minority describe something that needs a clinician’s attention today, and they are frequently not written as urgent, because the patient does not know that what they are describing matters. Processing the inbox in arrival order means the second category waits behind the first. Triaging it by hand means a clinician or a trained administrator reads every message, which is the correct approach and the reason inboxes fall behind.
Patient Message Triage triages an inbound message against the practice’s own policy, looks up the record, and drafts either an administrative reply or a clinical escalation. Every reply waits for a human. Two constraints define what this is: it triages against the practice’s stated policy rather than exercising clinical judgement of its own, and it drafts rather than sends. The agent’s contribution is that a message needing clinical attention is identified and escalated within minutes of arrival instead of hours, and the administrative reply arrives pre-drafted for someone to check. One agent covers this process, and deploying it is a clinical governance decision, not an IT one.
What this moves
- Clinical messages reaching a clinician promptly
- Messages needing clinical attention are separated from administrative ones against the practice's stated policy, rather than waiting in a single inbox by arrival order.
- Replies sent without a human reading them
- Every reply waits for a person, so no patient receives an answer that a clinician or administrator has not seen.
Healthcare
How AI agents handle patient communications
Drawn from the 1 agent above — what they require, how they run, and what comes back.
What they need
- Inbound patient message
What comes back
- Triage
- Safety flags
- Draft reply
- Patient record
- Policy sections used
- FHIR R4 bundle
How they run
- Triggered by an incoming message
- 1
- Steps per run
- 6
- Credits per run
- 5
- Use a knowledge base
- 1
Where patient communications fits in healthcare
Documentation takes a large share of every shift and is a leading reason people leave. Drafting it from the record is where the time comes back.
Next Step
Deploying patient communications agents
These run as-is against the inputs listed above. Most deployments adapt one — a different source system, a different tolerance, a different approval path. The first call establishes which.