All use cases

Upsell & Cross-sell

Second sale inside the care conversation (telehealth)

A patient on the care chat mentions they're constantly hungry on the program — the agent helps first, then offers the nutrition add-on that actually fits, never a pitch.

Next best thingClose the loopBehavioral triggerChatWhatsAppHealth & Telehealth

What it does

Telehealth care chats are full of patients at the exact moment they're thinking about their treatment: asking about side effects, checking on a refill, describing how a program is going. Most teams resolve the message and close the thread, and a genuinely fitting next step — a complementary supplement, a higher-touch coaching tier, a relevant test — goes unmentioned, leaving the care surface a pure cost center. This agent treats the care conversation as a compliant sales channel. Once the patient's question is genuinely handled, it surfaces the clinically-appropriate add-on drawn from what they asked, their program stage, and memory — with a one-line reason and never a word of medical advice it isn't allowed to give.

How it works

  1. 1
    Trigger. a care or coaching contact arrives on chat or WhatsApp — a side-effect question, a check-in on progress, a request that reveals an unmet need.
  2. 2
    Decision. clinical red-flag detection runs first; anything urgent escalates to a clinician instead of surfacing an offer. Then the agent resolves the care question, and only if a genuine, suitable need remains does it weigh a second sale — program stage, suitability rules, prior declines, and frequency caps all factor in, and a quality judge gates both tone and fit.
  3. 3
    Action. the offer lands inside the same conversation as a natural next turn — the supplement matched to a program-stage need, the coaching tier for someone struggling, the test that fits their question — each with a one-line reason and no clinical claim beyond approved language.
  4. 4
    Follow-through. accepted offers hand off to a secure checkout on the provider's hosted payment step; the agent never takes the payment itself. The outcome — taken, declined, or not raised — writes back to memory with a cool-down so the patient isn't nudged again next time, and lift is measured against a holdout.
Example
CustomerI'm three weeks in and the medication's fine, but I'm starving by mid-morning and losing track of what to eat.
AgentThat mid-morning hunger is one of the most common things people hit around week three. I can't change your dose here, but appetite and nutrition are exactly what our coaching add-on is built for — a coach who plans meals around your routine and goals. Want me to share the details?
CustomerYeah, how much is it?
AgentI'll lay out the options and hand you to our secure checkout if you'd like to go ahead — and we'll keep this thread right where we left it.

Configuration

How the agent is wired for this use case.

Triggeran inbound care-chat or WhatsApp message classified as a need_revealing_message (side-effect, progress, or unmet-need intent) after the service question is resolved.
Tools & actions
  • Clinical safety classifier · screen every message for red flags first; route anything urgent to a clinician before any commercial turn.
  • Care knowledge base · answer the care question within approved, non-prescriptive language.
  • Product / program catalog · read suitability rules and program stage to select a clinically-appropriate add-on (supplement, coaching tier, test).
  • Customer memory · read program stage, prior offers and declines; write back the outcome with a cool-down.
  • Messaging channel · surface the offer in-thread and hand off to the provider's hosted secure payment step.
Autonomyanswering the care question and surfacing a suitability-checked offer run unattended under judge gating on tone and fit; red-flag detection and clinical escalation are mandatory and non-bypassable; the agent gives no dosage or medical advice, and payment is a secure hand-off to the provider's hosted step, never taken in-conversation.
Channelschat · whatsapp
Escalationa clinical red flag, a symptom or dosage question beyond approved scope, a complaint, or a vulnerability signal hands off to a clinician or human with full conversation context.

What you need

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

Signals

new care-chat or WhatsApp contact, intent and red-flag classification, resolution-complete and unmet-need events.

Data

program stage and treatment context, suitability rules, add-on/program catalog, consent state, memory of prior offers and declines.

Guardrails

clinical red-flag detection runs before any commercial nudge and escalation is mandatory; no proactive dosage or medical advice from the agent; offer only after the care question is resolved and only on genuine suitability; judge gating on every offer; decline cool-downs and frequency caps; secure payment hand-off only, the provider keeps owning the channel.

Metrics it moves

  • attach-rateup, as suitability-checked add-ons ride inside care the patient is already engaged in.
  • revenue-per-conversationup; the care surface starts carrying attributable, compliant sales instead of only cost.
  • conversion-rateup versus broadcast promotions, because the offer arrives when the patient's own message justifies it.

Related use cases

Upsell & Cross-sell

Second sale inside the support conversation

the retail parent of this in-conversation second sale

Upsell & Cross-sell

WhatsApp care-opener upsell

the same care-first sale, but agent-initiated outbound

Upsell & Cross-sell

Free-to-paid nudge in the care chat

monetizing a free care tier in the same thread

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