All use cases

Engage / Lifecycle

Enrollment-assigned continuous GLP-1 coach

Not a help bot you summon when something breaks — a coach assigned the day you enroll, who stays with you across the whole program and remembers exactly where you are in it.

Get aheadTime-based triggerBehavioral triggerChatPushWhatsAppIn-appGLP-1 & PeptidesHealth & TelehealthSubscription

What it does

Most GLP-1 support waits to be asked: the patient has to notice a problem, open a ticket, and re-explain their whole history to whoever answers. By then the doubt has set in, and the known quit windows — the early plateau, the four-month dip, the silent maintenance drop — pass with no one reaching in. A reactive chatbot cannot hold a twelve-month course together, because it forgets the patient between every conversation. This is a coach assigned at enrollment and kept in continuous contact across the full program, carrying one memory of the current dose, the last weigh-in, the last side effect, and the weeks until the next titration step — so it reaches out at the milestones and the quiet signals that decide whether the patient stays, instead of waiting for a ticket that never comes.

How it works

  1. 1
    Trigger. the coach is assigned at enrollment and stays active for the program; it acts on milestones on the clock (an approaching titration step, a monthly mark), behavioral signals (a missed log, engagement going quiet), and risk signals from a connected wearable.
  2. 2
    Decision. it reads the patient's single program memory — dose history, last weigh-in, last side effect, weeks-to-next-titration, what was said last — and weighs whether reaching out now genuinely helps; every unprompted message passes a quality judge under quiet hours and frequency caps.
  3. 3
    Action. a short, personal touch on the channel the patient answers, pegged to where they are in treatment — anticipating the side effect they reported before a step-up, naming the plateau feeling at the window it tends to hit, marking a non-scale win when the scale stalls.
  4. 4
    Follow-through. every exchange writes back to the same memory, so the next clinician visit and the next nudge open on continuity, not from scratch. The window-specific saves and the symptom flows hang off this one memory; any clinical red flag is a hard handoff to the care team, never an automated reply.
Example
AgentYou're four months in this week — a stretch where a lot of people feel the scale slow down and start to wonder if it's still working. Looking at your last few weigh-ins, you're right on the curve we'd expect. How are you feeling about it?
CustomerHonestly, a bit flat. I was losing faster at the start.
AgentThat shift is normal around now, and your energy logs are actually up since last month. Let's protect the muscle while the scale catches up — I'll check in midweek, and if you want, I can have your care team look at where you are.

Configuration

How the agent is wired for this use case.

Triggerassignment at enrollment, then program-milestone events (titration step approaching, monthly mark) from the program backend, behavioral signals (missed log, engagement decay), and wearable risk signals.
Tools & actions
  • Program backend / EHR · read dose history, titration schedule, last weigh-in, and last side effect; write replies back to the patient record.
  • Memory layer · hold one continuous program memory across channels; write each exchange back so the next visit and nudge open on continuity.
  • Wearable / device feed · ingest weigh-ins, injection logs, and sleep or activity data as trigger and context.
  • Messaging channel · send the personal touch on the channel the patient answers (chat, push, WhatsApp, in-app).
  • Care-team escalation path · route any clinical red flag to a clinician.
Autonomyunprompted touches send unattended only after a quality judge clears them under quiet hours and frequency caps; the coach works within the program's protocol and the medication journey, gives no dose or medical advice, and treats any clinical red flag as a hard human handoff rather than an automated reply.
Channelschat · push · whatsapp · in-app
Escalationa concerning symptom, distress, or any clinical red flag in a reply escalates straight to the care team.

What you need

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

Signals

enrollment, program-schedule events (dose due, titration step), logging events (injection, weigh-in), engagement-decay signals, wearable data where connected.

Data

patient profile and goal, dose and titration history, last weigh-in and last side effect, the single cross-channel program memory, consent and quiet-hours preferences.

Guardrails

judge review of every unprompted message; the coach stays within protocol and improvises no medical or dose advice; clinical red-flag detection with a hard human handoff; per-channel consent, frequency caps, and stale-touch cancellation when the patient shows up first.

Metrics it moves

  • adherenceup, because the moments that break a course get a timely, personal touch from a coach who already knows the history.
  • churndown, the known quit windows are met by an assigned coach instead of passing in silence.
  • ltvup, continuous contact across the full program converts known quit-reasons into retained tenure.
  • dau-mauup, assigned touches open into real coaching conversations rather than dead-end notifications.

Related use cases

Engage / Lifecycle

Proactive health-coach companion

the cross-vertical companion this specialises into an enrollment-assigned, program-bound coach

Engage / Lifecycle

Between-doses lull check-in

the weekly-trough cadence that runs on this coach's memory

Retain / Save

Four-month drop-off re-engagement

one of the named windows this continuous coach is built to catch

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