All use cases

Upsell & Cross-sell

Complete-the-kit cross-sell

Your customer just built a tennis basket with no racket bag — the agent spots the gap and completes the kit before checkout.

Next best thingGet aheadBehavioral triggerChatIn-appE-commerce & Retail

What it does

A customer assembles a basket that is almost right: shoes and a polo, but no socks; a racket, but nothing to carry it in. They check out anyway, and the missing piece becomes someone else's sale next week. The agent reads live cart state, recognizes the obvious companion the basket is missing, and offers it with a one-line reason — "most buyers add this" — growing the basket without reaching for a discount.

How it works

  1. 1
    Trigger. the storefront's widget API streams live cart state; the agent detects a kit-gap pattern — items that almost always travel together, with one missing.
  2. 2
    Decision. the agent weighs the gap against the customer's purchase history and what it already knows about them (sizes, preferences), and a judge gates the nudge — it only goes out if it is genuinely useful, on-policy, and not a repeat.
  3. 3
    Action. a short in-chat or in-app suggestion naming the specific companion item and why it fits this exact basket, with a one-tap add. No discount attached — the relevance is the offer.
  4. 4
    Follow-through. the outcome (added, declined, ignored) is written back to memory so the next nudge is smarter; if the customer checks out or empties the cart first, the pending nudge is cancelled rather than sent stale.
Example
AgentOne more thing before you check out — most players who buy these shoes and a polo add a pair of court socks to round out the kit. Want me to drop a pair in your size into the basket?
CustomerSure, add them.
AgentDone — they're in your cart in your usual size. Your basket's ready whenever you are.

Configuration

How the agent is wired for this use case.

Triggercart_updated event streamed from the storefront's cart/widget API, surfacing a kit-gap pattern
Tools & actions
  • Storefront cart API · read live cart state; add the companion item on one-tap accept
  • Product catalog · resolve companion/affinity relationships and size availability
  • CRM / memory store · read purchase history and size preferences; write back the outcome
  • Messaging channel · render the in-chat or in-app suggestion with a one-tap add
Autonomythe nudge runs unattended behind a relevance/policy judge; the basket add executes only on explicit customer tap, and the client keeps owning the surface
Channelschat · in-app
Escalationno automated handoff; a declined or ignored nudge ends the flow and is logged to memory, and a pending nudge is cancelled if the customer checks out or empties the cart first

What you need

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

Signals

live cart state via the storefront widget API (cart_updated, items, quantities)

Data

product catalog with companion/affinity relationships, customer purchase history and size/preference memory

Guardrails

judge gating on every unprompted nudge; one nudge per session; suppression when the basket is already complete or the customer has declined the same add before; the client keeps owning the surface

Metrics it moves

  • attach-rateup, because the companion item is offered at the exact moment the basket reveals the gap
  • aovup, through full-price basket completion rather than discount-led add-ons

Related use cases

Upsell & Cross-sell

Next-best-product recommendation block

in-conversation sibling triggered by a stated need

Convert

Checkout-stall rescue

catches the same session when it stalls instead of growing

Convert

Back-in-stock ping with cross-sell

the same companion logic, triggered by inventory

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