All use cases

Collect / Recover

GLP-1 plan, pricing and billing conversation handler

"Why was I charged this?" and "what does my plan actually include?" are some of the most-asked questions a medication program gets — the agent answers them in the thread, in plain language, and only hands off when money moves.

Close the loopBehavioral triggerSystem triggerChatEmailWhatsAppIn-appGLP-1 & PeptidesHealth & TelehealthSubscription

What it does

Billing is one of the named recurring support-conversation types in a GLP-1 program, and it sits inside the large, unglamorous majority of tickets that nobody wants to staff. Patients ask why they were charged, what their plan tier includes, how the multi-month options compare, and how to change or pause a plan — and each one waits in a queue for a human to read an account and explain it. The agent resolves these plan, pricing, and billing questions directly: it verifies identity, pulls the account and charge history, explains the charge or plan in plain language, and makes the plan change the patient asks for — handing off to a secure payment step only when an actual payment is involved. Pure operating margin recovered from a conversation that used to cost a person's time.

How it works

  1. 1
    Trigger. a patient asks a plan, pricing, or billing question in chat, email, WhatsApp, or in-app — "what am I paying for", "why this charge", "can I switch to the longer plan", "pause my billing" — or a billing-event message lands that the patient is querying.
  2. 2
    Decision. the agent verifies identity, retrieves the account, the active plan tier, and the charge history, and reads the program's billing policy for what it may explain and which plan changes it may make without a human.
  3. 3
    Action. it explains the charge or the plan options in plain language in the same thread, and where the patient wants a permitted change — switch tier, change cadence, pause — it makes it against policy and confirms the new state. Anything touching an actual payment routes to the program's secure hosted step.
  4. 4
    Follow-through. the change and its confirmation are written back to billing so the account state is never ambiguous; a charge dispute or a refund request escalates to a human with the account context attached; a clinical question riding along on a billing thread is routed out, not answered as advice.
Example
CustomerI got charged again and I thought I was on the longer plan. Can you explain?
AgentYou're verified. You're on the monthly plan, so this is your regular monthly charge — the longer plan bills once up front and works out lower per month. Want me to switch you to it from your next cycle?
CustomerYes please, the longer one.
AgentDone — you'll move to the multi-month plan at your next renewal, and I've noted it on your account. When the first payment for it comes up I'll take you to the secure checkout; nothing changes on today's charge.

Configuration

How the agent is wired for this use case.

Triggera plan/pricing/billing question detected in chat, email, WhatsApp, or in-app, or a billing-event message the patient is querying.
Tools & actions
  • Billing platform · verify identity, retrieve the account, plan tier, and charge history; explain a charge; make a permitted plan change and write the new state back.
  • Knowledge base · source the plan-tier definitions and billing-policy explanations the agent gives in plain language.
  • Messaging channel · handle the conversation and return the confirmed plan state to the thread.
  • Payment provider · hand off to the program's secure hosted step whenever an actual payment is involved; the agent never processes the payment itself.
  • CRM · log the question type and resolution; route a dispute or refund to a human with context.
Autonomyidentity verification, charge explanation, and policy-permitted plan changes run unattended within the program's billing policy; any actual payment completes only on the secure hosted step, and confirmation is read from the system of record, never assumed. A charge dispute, a refund, or a hardship signal escalates to a human.
Channelschat · email · whatsapp · in-app
Escalationa charge dispute, a refund request, a hardship signal, or any plan change outside policy escalates to a human; a clinical question on a billing thread routes to the clinical path.

What you need

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

Signals

plan/pricing/billing-question intent across chat, email, WhatsApp, and in-app, plus billing lifecycle events (charge_posted, plan_renewed) the patient may query.

Data

the account and plan-tier structure, charge and invoice history, the plan-change and refund policy, identity-verification fields, consent state.

Guardrails

identity verification before any account detail is shown; plan changes strictly within program policy; any payment completed only on the secure hosted step; confirmation read from the system of record, never assumed; disputes, refunds, and hardship escalate to a human; no clinical or dosage advice on a billing thread.

Metrics it moves

  • ticket-deflectionup, by resolving the named, high-volume plan-and-billing question category in the thread instead of in a human queue.
  • recovered-revenueprotected, because plan confusion and silent charge questions are resolved into retained, correctly-tiered plans rather than cancellations.
  • time-to-resolutiona charge explained or a plan changed in one thread at any hour, instead of an email round-trip or a callback.

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