> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trychannel3.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Conversational Shopping

Building a conversational shopping assistant? The Conversational Agent can power it, from chat to search.

## A shopping conversation

Here's what a typical session looks like.

<Steps>
  <Step title="Shopper asks for products">
    > **Shopper:** I need waterproof hiking boots for day hikes, under \$200

    **What happens:** Channel3 creates a new conversation (`conv_123…`), searches the catalog, and streams back:

    * A short assistant reply ("Let me pull up some options…")
    * Product cards with titles, images, prices, and buy links
  </Step>

  <Step title="Shopper refines">
    > **Shopper:** in brown instead

    **What happens:** You send only the new message plus the saved `conv_123…` ID. Channel3 remembers the conversation, and runs a **new search** for brown options.
  </Step>

  <Step title="Shopper asks about what's on screen">
    > **Shopper:** Which of those is better for wide feet?

    **What happens:** The assistant compares the products already in the conversation — no new catalog search. It may highlight one of the cards already shown.
  </Step>

  <Step title="Page refresh">
    The shopper comes back. Fetch the saved conversation to reload the full message history — user messages and assistant replies with product cards — and rebuild your chat UI.
  </Step>
</Steps>

## Build it

See the [Conversational Agent quickstart](/conversations/quickstart) for minting a client token, streaming events, and how to render product cards in your UI.