All use cases

Win back

Regain-triggered win-back

The patient who stopped months ago just logged that the hunger is back. That's the moment to reach out — not a date on a win-back calendar.

Get aheadTime-based triggerBehavioral triggerRisk triggerEmailPushWhatsAppChatGLP-1 & PeptidesHealth & TelehealthSubscription

What it does

Most patients regain weight within a year of stopping a GLP-1 medication, and the first sign is usually behavioural — returning food noise, appetite creeping back — well before the scale moves or a renewal date arrives. Calendar-based win-back blasts miss this window entirely, reaching out on a fixed timer that has nothing to do with what the patient is actually feeling. The agent watches for the regain signal in a lapsed patient and reaches back with a compliant, memory-personalized message at the moment it matters, routing the restart-or-re-titrate decision to a clinician rather than implying a dose. The reactivation lands as relevant care instead of a generic "we miss you."

How it works

  1. 1
    Trigger. a lapsed patient produces a regain signal — re-engagement in the app, self-reported returning food noise or appetite — optionally combined with time since stopping, rather than a lapse timer alone.
  2. 2
    Decision. the agent loads memory (last dose, why they stopped, what worked, consent and channel preference), confirms this reads as genuine regain and not noise, and a quality judge gates the message for relevance, tone, and compliance; the restart decision is framed as a clinician's, never the agent's.
  3. 3
    Action. a personal, specific message goes out on the patient's preferred channel (email, push, WhatsApp, or chat) — naming the signal and offering a safe path back — and opens a two-way conversation that leads to a clinician review for the restart, with a secure payment hand-off for any commercial step.
  4. 4
    Follow-through. replies are handled in-dialog through to a clinician handoff or a clear no; the outcome is written back to memory; if the patient re-enrolls on their own first, pending outreach is cancelled rather than sent stale, and reactivation is measured per cohort against a holdout.
Example
AgentHi — I noticed you mentioned the hunger creeping back this week. That's really common after stopping, and it doesn't mean you're back to square one. If you'd like, I can set up a quick review with a clinician to talk through restarting safely.
PatientYeah, the cravings are definitely back. Can I just go back on the same dose?
AgentThat's exactly the call your clinician should make, since the right restart depends on how long it's been — I won't guess at a dose. I'll book the review and bring your history across so you're not starting from scratch.

Configuration

How the agent is wired for this use case.

Triggera regain_signal_detected behavioural event in a lapsed patient — app re-engagement or self-reported returning appetite or food noise — optionally with time-since-stop.
Tools & actions
  • Persistent memory · read last dose, stop reason, what worked, consent and channel preference; write the outcome back
  • App backend · detect re-engagement and self-reported appetite or food-noise signals
  • CRM · read lapse and lifecycle state, suppress do-not-contact, update reactivation status
  • Messaging channel · send the memory-personalized two-way message on the patient's preferred channel
  • Scheduling system · book the clinician restart review with history attached
  • Secure payment hand-off · complete any commercial step through the client's hosted step, never in-chat
Autonomytiming, signal confirmation, and the outbound send run unattended only after a judge clears the message for relevance, tone, and compliance, with suppression lists, frequency caps, and a per-cohort holdout applied; the restart or re-titration decision is a mandatory clinician hand-off, never an automated reply; money moves only through a secure payment hand-off.
Channelsemail · push · whatsapp · chat
Escalationthe restart or dose-change decision is routed to a clinician for review; the agent never advises a dose, a schedule, or a self-restart.

What you need

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

Signals

regain indicators in a lapsed cohort (app re-engagement, self-reported returning food noise or appetite), time since stopping.

Data

persistent patient memory (last dose, stop reason, what worked, preferences), lapse and lifecycle state, channel consent.

Guardrails

the restart or re-titration decision always routes to a clinician — no dose, schedule, or self-restart advice; judge gating on every unprompted message; suppression of do-not-contact and recently contacted patients; frequency caps and instant opt-out; stale-outreach cancellation on organic return; per-cohort holdout; payment only via secure hand-off.

Metrics it moves

  • reactivation-rateup: the message lands on the regain signal, the moment a stopped patient is most open to restarting. The category context — most patients regain a large share of lost weight within a year of stopping — is what makes this window worth catching, as proof of the problem rather than a Mentiora result.
  • save-rateup: reaching at the appetite signal catches patients before regain compounds.
  • recovered-revenueup, tracked per cohort from outreach to a clinician-approved restart against a holdout.
  • ltvup: a safe, well-timed restart extends the patient relationship rather than ending it at the lapse.

Related use cases

Win back

Memory-personalized win-back

the cross-industry win-back motion this specialises to the regain signal

Retain / Save

Off-ramp safe-transition agent

the discontinuation flow whose nudge-back path this card fires

Retain / Save

Food-noise early-warning save

the same appetite signal used to retain before the patient stops

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