our process

How Shoppa takes you from catalogue to live agent

It starts with your product data, because your agent is only as good as what it knows. Then the embed puts the agent on your site, and the conversation does the rest.

01

Connect

Point Shoppa at the feed you already have

Shoppa ingests your product feed - JSON or CSV - and turns it into the knowledge your agent sells from. Titles, descriptions, categories, prices, sizes, colours, stock, and images: the fields you already maintain are all it needs.

Updates are incremental: re-ingest the feed and only changed products are touched. Mark a product out of stock and it leaves the agent’s in-stock answers - your agent never recommends what you can’t sell.

Included in this phase

  • Product Feed
  • JSON or CSV
  • Prices in AUD
  • Sizes & Colours
  • Stock & Availability
  • Images

02

Embed

Two lines of code, in your brand

The complete integration is two script tags. Drop them on your site, point them at your retailer ID, and the agent appears in your storefront - designed to sit inside your brand rather than a bolted-on chat box.

There’s nothing to rebuild and nothing to migrate: your site stays exactly as it is, and the agent runs on top of it.

<script src="https://api.shoppa.au/embed/shoppa.js"></script>
<script>Shoppa.init({ retailerId: "yourstore", apiBase: "https://api.shoppa.au" })</script>

03

Sell

One conversation, start to finish

Your agent runs the whole journey: product discovery, add to cart, checkout in the chat, and post-sale order status. Customers ask in their own words and get answers grounded in your catalogue - with the sale and the relationship staying on your site.

After the sale, the same conversation keeps working: ‘where’s my order?’ gets a real answer, tied to the order the agent just took.

our values

Built for retail conversations, not generic AI adoption

Putting a trustworthy agent between you and your customer is a technical and commercial problem. These principles keep Shoppa focused on answer quality and revenue rather than AI theatre.

Agent-readable data

Product records need complete identifiers, attributes, descriptions, and categories that agents can compare without guessing.

Freshness as a signal

Stock, price, availability, and status changes must reach your agent before stale data costs a sale.

Grounded answers

Your agent answers from your catalogue, not from the open internet. If it isn’t in your data, it isn’t in the answer.

Brand-safe conversations

The agent lives in your brand, on your site, speaking to your customers - never a third-party surface between you.

Ownership

The conversation, the checkout, and the customer relationship stay yours. That is the point.

One conversation

Discovery, checkout, and order support belong in the same thread, not three different tools.

Put your own agent in the conversation

The retailers deploying their own shopping agents today are building customer relationships that compound tomorrow.

Contact us