All use cases

Retain / Save

Safe off-ramp companion for hormone therapy

The patient who decides to stop testosterone or HRT usually just disappears — and an unsupervised stop is the one moment that carries real clinical risk.

Close the loopGet aheadBehavioral triggerRisk triggerChatIn-appPushEmailGLP-1 & PeptidesHealth & TelehealthSubscription

What it does

When a hormone-therapy patient decides to come off, most programs find out only when the subscription lapses, so the patient churns silently and an abrupt unsupported stop goes unmanaged. That is the wrong moment to be absent: stopping testosterone suddenly can crash a patient who has shut down their own production, and stopping HRT carries its own bone-health considerations. The reasons differ too — fertility plans, side-effects, and cost each need a different answer, not one generic save offer. The agent catches the stop signal, flags it to the clinician for a supervised taper rather than a hard stop, offers a maintenance or lighter-touch path where it fits, and leaves a clean door open for return.

How it works

  1. 1
    Trigger. the patient signals stopping — a "thinking of coming off" message in chat, a fertility plan mentioned on testosterone, a cancel-flow entry, or a refill that lapses with engagement decay.
  2. 2
    Decision. the agent reads the patient's memory (therapy type, time on treatment, stated reason) and classifies the stop-reason into its branch — fertility, side-effects, or cost — while recognizing that the safe wind-down itself is a clinical decision it must route, never script.
  3. 3
    Action. on the patient's consented channel the agent runs the reason-matched branch, surfaces a maintenance or lower-touch option where one is clinically appropriate, and hands the taper decision to the clinician with the patient's context attached. It never advises a dose or a schedule.
  4. 4
    Follow-through. the reason and outcome write back to memory; if the patient stays, queued cancel-flow outreach is cancelled; if they leave, a nudge-back path stays armed for a clinician-led return. Save impact is measured against a holdout.
Example
AgentYou mentioned wanting to come off testosterone — that's your call, and I want to make sure it's done safely rather than just stopped, since coming off abruptly can leave you feeling worse before it settles. Can I ask what's behind it: starting a family, how you're feeling on it, or the cost?
CustomerWe're trying for a baby, so I think I need to stop.
AgentThat's a common and good reason to change the plan, and it's exactly the kind of thing your clinician should guide rather than a hard stop on your own. I'll flag this to them now with your history so they can set up a safe wind-down and talk through fertility options — and I'll keep your account ready if you want to pick treatment back up later.

Configuration

How the agent is wired for this use case.

Triggera stop signal — a discontinuation-intent phrase in chat, a fertility-plan mention on testosterone, a cancel-flow entry, or a lapsed refill with engagement decay — sourced from the program app, the subscription billing system, and the messaging channel.
Tools & actions
  • CRM / customer record · read therapy type, time on treatment, and stated reason; classify the stop-reason; write reason and outcome back
  • Subscription billing · read plan and lifecycle state; surface a maintenance or lower-touch tier where clinically appropriate
  • Knowledge base · retrieve reason-specific, non-prescriptive context on why a supervised wind-down matters
  • Clinical / EHR system · route the taper decision and any fertility or abrupt-stop signal to the clinician with full context
  • Messaging channel · run the branched off-ramp conversation and the return-path follow-up
Autonomyreason-classification, option-surfacing, and outreach run unattended behind judge gating; this is a detect-and-escalate clinical posture — the agent never sets, changes, or schedules a dose, and the wind-down itself always hands to a clinician under the clinical-safety policy class.
Channelschat · in-app · push · email
Escalationany taper or wind-down decision, a fertility plan on testosterone, or a clinically unsafe abrupt-stop intent hands off to a clinician.

What you need

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

Signals

discontinuation-intent phrases, fertility-plan mentions, cancel-flow entry, refill lapse, engagement-decay scores

Data

therapy type and time on treatment, stated stop-reason, plan and maintenance options, channel consent, off-ramp memory

Guardrails

judge gating on every unprompted message; detect-and-escalate only, no dose or taper-schedule advice; fertility and abrupt-stop signals hard-escalate to a clinician; frequency caps; stale-outreach cancellation on an organic stay; holdout measurement of saves

Metrics it moves

  • save-rateup, by meeting the stop decision with a reason-matched maintenance or lighter-touch path instead of a silent lapse
  • churndown, as cost and side-effect stops convert into supervised continuations rather than hard exits
  • reactivation-rateup, because the armed return path catches patients whose reason to stop later resolves

Related use cases

Retain / Save

Early-discontinuation save for new hormone starters

the front-loaded quit; this card handles the considered exit

Prevent / Care

Hormone lab-result dose-review handoff

the same clinician-handoff posture on a results trigger

Retain / Save

Off-ramp safe-transition agent

the GLP-1 regain-driven off-ramp this is the hormone counterpart to

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