> For the complete documentation index, see [llms.txt](https://alhena.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://alhena.gitbook.io/docs/features/ai-search.md). # AI Search AI Search is a full-page product search for your website. Shoppers describe what they want in their own words and get a results page with the right filters already set. AI Search turns a shopper's sentence into a filtered product results page. A shopper who types *"navy suit for a summer wedding under $800"* lands on navy suits under $800, with those filters shown at the top of the page and ready to change. It works like your store's search page, except the shopper never has to find and tick the filters themselves. AI Search and [Conversational Search](/docs/features/conversational-search.md) open from the same code on your website. Conversational Search answers in a chat conversation. AI Search shows a product grid with filters. Once AI Search is turned on for your store, it takes the place of Conversational Search.

AI Search results for "hydrating face care for dry winter skin under $90", with the filters it read from the search and suggested filters below them

## Key Features * **Understands full sentences.** AI Search reads the product type, colors, materials, patterns and price from what the shopper wrote, and applies them as filters. Each filter shows at the top of the page, where the shopper can remove or change it. * **Suggested filters.** When a shopper describes an occasion or a use, such as *"suit for a summer wedding"*, AI Search suggests filters that suit it (for example, linen) without applying them. The shopper adds the ones they like with one click. * **Close matches instead of an empty page.** When no product matches every filter, AI Search shows the closest products and says which filter it dropped. **Show exact matches only** puts that filter back. * **Filters built from your catalog.** The filter panel shows your own categories and product attributes, a price range the shopper can narrow, and customer rating when your store has reviews. When a shopper asks for something cheap or high-end, AI Search applies a budget, mid-range or premium price level, which the shopper can remove. * **Sorting.** Best match, price low to high, or price high to low. * **Works on mobile.** On small screens the filters sit above the results. ## How It Works 1. The shopper types what they're looking for into your search bar or the AI Search page and presses Enter. 2. AI Search reads the sentence, picks the matching categories and filters, and shows the products that match. 3. The shopper can remove a filter, add a suggested one, change the price, or sort, and the results update. **Reset to AI** brings back the filters AI Search chose. 4. Results load 12 at a time, with a **Show 12 more** button for the next ones. Clicking a product opens its page in a new tab, so the search stays open. ## Getting Started 1. **Ask Alhena to turn on AI Search.** Our team prepares your product catalog for AI Search, learning its categories and product attributes, and then switches your store to AI Search. 2. **Install on your website.** Follow the [Installation](/docs/features/ai-search/installation.md) guide. If Conversational Search is already on your site, there is nothing to change. 3. **Try it.** Search your store in your own words and check the results and filters. {% hint style="info" %} The **Customize** settings for AI Search in the dashboard (start page heading, starter questions and colors) apply to Conversational Search. The AI Search results page uses its own layout. {% endhint %} ## Related * [Installation](/docs/features/ai-search/installation.md): Add AI Search to your website * [Search Box Suggestions](/docs/developer-reference/website-sdk/examples/search-suggestions.md): Show suggestions in your own search box as the shopper types * [AI Search API](/docs/developer-reference/api/ai-search.md): Build your own search experience on the same engine * [Conversational Search](/docs/features/conversational-search.md): The chat-style search experience