Integration with a product feed (XML, Google Merchant)

Connect any online store to your ChatLab chatbot through a product feed. If your platform has no native ChatLab integration, the feed is the way to give the chatbot your catalog: it reads the products from the feed, indexes them with AI Search and answers with product cards that link to your store.

Important: E-Commerce Best Practices

After completing the integration below, read the comprehensive guide on optimizing your chatbot for e-commerce: How to set up ChatLab for your E-commerce store

What the Integration Provides

AI product search -- Customers describe what they need in their own words and the chatbot finds matching products in your catalog. The answer contains:

• Product name and description
• Price (and sale price when the feed has one)
• Product image
• Direct link to the product page in your store
• Brand, availability, GTIN / SKU and other attributes from the feed

Categories -- When the feed carries the product type (for example "Home > Kitchen > Kettles"), the chatbot can narrow the search to a category.

Scheduled refresh -- The catalog is refreshed on the schedule you set, so new products, prices and availability reach the chatbot without your intervention.

What the Integration Does Not Provide

The feed is a file, not a connection to your store, so the chatbot cannot:

• check the status of an order
• add products to the cart or place an order
• read live stock levels between refreshes

Product cards always link to your store, where the customer finishes the purchase. If you need order tracking or add to cart, use a native integration.

Requirements

• The Feed XML integration included in your plan
• The AI Search product pack assigned to the chatbot: it decides how many products can be indexed
• A product feed in the Google Merchant (Google Shopping) XML format, available at a public https address. Compressed feeds (.gz) and feeds protected with a username and password are supported.

Step 1: Get the Product Feed From Your Store

Most platforms export a Google Merchant feed out of the box or through a free extension:

• Magento / Adobe Commerce -- a Google Shopping feed extension (for example from the Magento Marketplace) generates the XML at an address such as https://yourstore.com/feeds/google.xml
• IdoSell -- Marketing and integrations, Price comparison sites and marketplaces, Google Merchant: enable the feed and copy its address
• OpenCart -- Extensions, Feeds, Google Base: enable it and copy the data feed URL
• BigCommerce -- Channel Manager, Google Shopping, or a feed app from the BigCommerce App Marketplace
• Shopware -- Marketing, Product comparisons, Google Shopping: create the export and copy its URL
• osCommerce and custom stores -- any module that writes an RSS 2.0 or Atom feed with the g: product attributes

The feed must contain at least a product identifier (g:id), a name and a link. Price, image, description, availability and product type make the answers much better.

Step 2: Connect the Feed in ChatLab

1. Log in to ChatLab and select your bot
2. Go to the Connect tab
3. Click Connect on the Feed XML tile
4. Paste the Feed URL
5. If the feed requires a login, turn on The feed requires a username and password and enter the credentials
6. Optionally set the Product language (two-letter code of the language used in the feed)
7. Click Test feed -- ChatLab downloads the beginning of the feed and shows the format, the number of products found, the first products as they will be indexed and any warnings (missing prices, links, images or categories)
8. Click Connect

After a successful connection ChatLab starts indexing the catalog in the background. Depending on the size of the feed this takes from a minute to a few dozen minutes. You can follow the progress in Training > E-commerce catalog.

Step 3: Non-Standard Feeds (Field Mapping)

If your feed is not in the Google Merchant format, ChatLab can still read it:

1. Run Test feed first -- the element names found in your file appear in the mapping selects
2. Open Field mapping for non-standard feeds
3. If one product is not an item or entry element, enter the name of the Product element
4. Pick the element for each field: Product identifier (required), Product name, Description, Product page URL, Image URL, Price, Category path and the rest
5. Run Test feed again and check the sample products
6. Click Connect

Fields left on "Default (Google Merchant)" keep the standard names, so a feed that follows the standard needs no mapping at all.

Step 4: Refresh Schedule

1. Go to Training > E-commerce catalog
2. On the Feed XML card click Retrain
3. Retrain now! refreshes the catalog immediately
4. To refresh automatically, choose the frequency (every day, once a week or once a month), the day and the time and click Schedule retrain

Only products that changed since the last refresh are re-indexed, and products removed from the feed are removed from the chatbot's catalog. Scheduled refresh follows your plan: weekly and monthly schedules need the scheduled retrain feature, daily schedules need the daily retrain feature.

Step 5: Configure the AI Action

After connecting, the Feed XML - AI product search action is enabled automatically for chatbots with the AI Search product. To adjust it, go to Settings > Actions and filter by the Feed XML platform. You can change the action label, the instructions that guide the AI and turn on Include product card to give the chatbot the full product card instead of the short description.

Testing Your Integration

Open your chatbot widget and try:

• "Show me kettles under 50 dollars" (tests search with a price filter)
• "Do you have blue cotton t-shirts?" (tests attribute matching)

The answers should show products from your feed with links to your store.

Tips

• Put the product type in the feed (g:product_type) as a full path, for example "Clothing > Men > T-Shirts" -- it becomes the category the chatbot can filter by
• Keep descriptions in the feed complete; the chatbot uses them to match customer questions
• Variants of one product should share an item_group_id -- ChatLab shows them as one product with all colours and sizes
• If your feed is large, a compressed (.gz) feed downloads faster

Install Chatbot on Your Store

To add the chat widget to your storefront:

1. In ChatLab, go to the Deploy tab for your bot
2. Copy the code snippet
3. Paste it into the template of your store, before the closing </body> tag, or use your platform's "custom HTML" or "scripts" setting