All use cases

Prevent / Care

Async message-urgency triage for the clinician inbox

A night's worth of patient messages lands in one flat queue, and the one that says "sharp pain since my dose" is buried behind a run of shipping questions. The agent re-orders the inbox so the clinician opens the urgent one first.

Close the loopSystem triggerRisk triggerEmailIn-appGLP-1 & PeptidesHealth & Telehealth

What it does

A scaling GLP-1 program turns the clinician inbox into a wall: refill questions, billing notes, "is this normal?", and the occasional genuine red flag all arrive in the order they were sent, at the same priority. The clinician works top to bottom, so response time on the message that actually matters depends on luck, and as volume climbs the urgent case waits longer behind the routine. The agent reads each inbound patient message as it arrives, classifies it by clinical urgency against the program's policy, and re-orders the clinician's existing queue so the highest-acuity messages surface first with the routine ones beneath. It ranks and orders the queue; it does not reply to the patient, diagnose, or touch a dose.

How it works

  1. 1
    Trigger. an inbound patient message lands in the clinician inbox (via the messaging channel or care platform), or a batch arrives overnight for the morning queue.
  2. 2
    Decision. the agent reads each message, scores it against the program's clinical-acuity policy and the patient's memory (medication, titration stage, recent reports), and assigns an urgency band — with conservative defaults that round any ambiguity upward, so a possible red flag is never buried.
  3. 3
    Action. the inbox is re-ordered so the highest-acuity messages sit at the top, each tagged with the reason it was ranked there and a one-line context summary; routine messages keep their place beneath, nothing is hidden or auto-answered.
  4. 4
    Follow-through. the ranking writes back to the care record for audit, a message that reads as acute is surfaced for immediate attention rather than left in the daily queue, and a "this got worse" follow-up re-ranks instantly.
Example
Agent (in the clinician's inbox)I've re-ordered this morning's queue by urgency. Top of the queue: a patient reporting sharp abdominal pain radiating to the back since yesterday's dose — flagged urgent, full thread and medication history attached. The next few are titration questions; shipping and billing notes are sorted to the bottom.
ClinicianOpening the first one now.
AgentNoted — I'll keep the rest sorted and re-rank if anything new comes in or a patient says it's worsening.

Configuration

How the agent is wired for this use case.

Triggeran inbound patient message arriving in the clinician inbox (message_received) from the messaging channel or care platform, or a batched overnight queue assembled for the morning
Tools & actions
  • Messaging channel · read each inbound patient message as it arrives
  • Clinical/EHR system · pull the patient's medication, titration stage, and recent reports to weigh the message in context
  • Knowledge base · apply the program's clinical-acuity policy and the per-category urgency criteria
  • Clinician inbox / care platform · re-order the queue, tag each message with its urgency band and a one-line reason, surface acute items for immediate attention
  • Care-record system · log the ranking and the routing decision for audit
Autonomymessage classification and queue re-ordering run unattended under a clinical-triage policy with conservative defaults — when unsure, rank higher. The agent orders the clinician's queue only; it never replies to the patient, diagnoses, or advises a dose or schedule, and every ranking is auditable. The clinician owns every reply and every clinical decision.
Channelsemail · in-app
Escalationa message scored acute is surfaced for immediate clinician attention rather than left in the daily queue; the agent never closes, answers, or resolves a clinical message itself — a human clinician acts on every one.

What you need

The inputs this use case runs on. Your channels stay yours; the agent supplies the judgment.

Signals

inbound patient messages (message_received), batched overnight queues, patient follow-up replies

Data

message text, patient medication and titration stage, recent symptom and visit history, persistent memory, the program's clinical-acuity policy and per-category urgency criteria, consent state

Guardrails

the agent ranks and orders only — no patient reply, diagnosis, dosing, or schedule advice ever; conservative defaults that round ambiguity upward so a possible red flag is never buried; clinician owns every reply and decision; full audit trail of every ranking; a worsening follow-up re-ranks immediately

Metrics it moves

  • time-to-resolutiondown, because the urgent message is read first instead of waiting its turn behind routine traffic in a flat queue
  • clinician-hours-savedup, as clinicians stop scanning a flat inbox to find what matters and open a queue already ordered by acuity
  • safe-escalation-rateup, since a buried acute message is reliably surfaced to the top rather than reached by chance partway through the day

Related use cases

Prevent / Care

Symptom-triage urgency-ranking layer

the patient-facing triage that ranks one message; this orders the whole clinician inbox

Prevent / Care

Doctor copilot — pre-drafted clinical reply with citations

what the clinician reaches for once an urgent message is on top

Prevent / Care

Clinician copilot — approve/decline decision packet

the decision-only handoff that follows once a message is opened

See it on your own customer journey

Bring one drop-off, one churn cliff, or one silent segment. We will show you what a proactive agent with memory and judgment does with it.

Book a demo