All use cases

Retain / Save

Early-discontinuation save for new hormone starters

A new hormone patient goes quiet two weeks after their first shipment, convinced it isn't working — the agent reaches in with the real timeline before they quit on a hunch.

Get aheadBehavioral triggerTime-based triggerRisk triggerPushIn-appEmailChatGLP-1 & PeptidesHealth & TelehealthSubscription

What it does

Hormone therapy has a front-loaded drop-off: the side-effects of starting often arrive before the benefit does, so a patient in the first weeks concludes "this isn't working" and stops before a fair trial. The signals are quiet — a "should I even bother" message, silence after the first shipment, an early cancel intent — and a program with no proactive layer only learns about it at the failed renewal. The agent watches for that early-quit signal in the starter window, reaches with a realistic "here is the normal timeline" reframe so the patient knows what to expect and when, and if the concern is genuine offers a clinician review rather than a pep talk. It saves the start without ever promising a result or adjusting a regimen.

How it works

  1. 1
    Trigger. an early-discontinuation signal in the first weeks of a new hormone Rx — a "is this working" message, engagement decay after the first shipment, or an early cancel intent.
  2. 2
    Decision. the agent checks where the patient sits against the program's expected symptom-relief timeline and their own logged experience, separating "too early to judge" from "side-effects without any benefit, worth a clinician's eyes."
  3. 3
    Action. it sends a timeline-grounded reframe on the patient's channel — what is normal this early, when relief typically lands — and where the picture warrants it, offers a clinician review instead of reassurance.
  4. 4
    Follow-through. the touch and the patient's reply write back to memory; a confirmed clinician review routes through the care team; if the patient re-engages on their own first, the queued nudge is cancelled.
Example
AgentYou started two weeks ago and I noticed things have gone quiet. This early, a lot of people feel the side-effects before they feel the benefit — for most, relief lands a little further in, not on day one. Want me to walk you through what's typical, or set up a quick clinician review if it really doesn't feel right?
PatientHonestly I figured it just wasn't doing anything.
AgentThat's the most common reason people stop too soon. Here's the realistic timeline for what you're on — and if you're still not feeling it after that, I'll book the review so your clinician can take a proper look.

Configuration

How the agent is wired for this use case.

Triggeran early-discontinuation signal in the starter window — an "is this working" message, post-first-shipment engagement decay, or early cancel intent, sourced from the messaging channel and the app backend
Tools & actions
  • App backend · detect the early-quit signal (engagement decay, cancel intent) inside the first-weeks window
  • Knowledge base · pull the program's expected symptom-relief timeline for the patient's therapy
  • Messaging channel · send the timeline-grounded reframe on push, in-app, email, or chat
  • Scheduling system · book a clinician review when the concern is genuine, not just early
  • Memory store · read the patient's start date and logged experience, write back the touch and outcome
Autonomythe reframe runs unattended behind judge gating, bounded to expectation-setting and timeline education; it never promises a result, diagnoses, or adjusts a regimen. A genuine "not working" picture is a clinician review, not an agent reassurance — detect, then escalate; no dose advice.
Channelspush · in-app · email · chat
Escalationa patient whose side-effects or lack of benefit warrant medical input hands to a clinician review through the care team rather than getting a coaching message

What you need

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

Signals

start date and starter-window flag, post-shipment engagement decay, early cancel intent, "is this working" messages

Data

the program's expected symptom-relief timeline, the patient's regimen and logged experience, persistent memory, channel consent

Guardrails

outreach bounded to expectation-setting and timeline education; no promised result, no diagnosis, no regimen change; detect-then-escalate for genuine concerns; judge gating on every message; frequency caps and one-follow-up rule; stale-nudge cancellation on self-serve re-engagement

Metrics it moves

  • save-rateup, by catching the quit-on-a-hunch before a fair trial in the front-loaded drop-off window
  • churndown across the vulnerable starter weeks where hormone programs lose patients before the benefit lands
  • adherenceup, as patients stay on long enough to reach the point where relief typically arrives

Related use cases

Retain / Save

Month-4 cliff companion coach (GLP-1)

the same save logic at the GLP-1 month-4 window, a different segment and timeline

Engage / Lifecycle

Hormone-symptom titration check-in

the scheduled symptom tracking the save can route into

Retain / Save

Churn-risk early-warning outreach

the general early-warning sibling this specializes for hormone starters

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