How it works for your shoppers#
Camille runs a French fashion store. A shopper types “echarpe” (scarf) into the search bar. The catalog spells it “écharpe” with the accent — a traditional search engine would return zero results, and the shopper would leave. CartAmplify Search matches anyway, returns the right products, and saves the sale.
A US electronics store sells “earbuds.” A shopper types “ear buds” (with a space). Or “earphones.” Or even “earpods.” CartAmplify understands all three are the same intent and returns the same relevant products.
Two shoppers search “shirt” on the same store. One has been browsing men’s formalwear for weeks. The other has only ever looked at women’s casual. They each see the same products — but ordered differently, prioritizing what each is more likely to click.
How CartAmplify learns and improves#
Three things happen, layered on top of each other:
1. Language understanding (active from day one). Every product title and description is read by a language model that captures meaning, not just letters. So “écharpe”, “echarpe”, and “scarf” all sit close together in the system’s mental map.
2. Popularity feedback (active after a few days of traffic). The system watches which results get clicked and which get ignored. Products people actually engage with rise; products people scroll past sink — automatically.
3. Per-shopper personalization (active after about a week of normal traffic). Once enough events have accumulated, the system trains a model that reshapes results for each visitor based on their browsing history.
What you can control#
You configure search behavior per language in Dashboard → AI Features → Search AI.
- Maximum results per page — How many products show before paginating.
- Exclude out-of-stock by default — Recommended on; toggle off for “notify when back in stock” use cases.
- Facets to show — Which filters appear in the sidebar (brand, size, color, price range, custom attributes).
- Default sort order — Relevance (the default) or any of Price, Popularity, Rating, Most-reviewed.
You don’t tune the AI directly — there are no relevance sliders, no scoring weights to learn. The system tunes itself based on what your shoppers do.
Multiple languages#
If you sell in more than one language, search is scoped per language. A shopper on your English storefront only sees English products; a shopper on the French site only sees French. Each language has its own search behavior and its own personalization model — they learn independently.
Setting it up#
If you’re on a supported platform (Shopify, WooCommerce, BigCommerce, Magento, OpenCart), the plugin wires the search bar and results page to CartAmplify automatically — no code on your side.
For custom storefronts or headless setups, your code calls POST /v1/search directly with the shopper’s query, then renders the returned product IDs in your own results UI. See the API reference for the request shape.