AI Product Recommendations

• What it does Recommends products from your catalog based on what the visitor needs
• Catalog sources Google Merchant feed, T-Soft, Shopify, WooCommerce
• Measurement Revenue from recommendations, with SKU-based attribution
• Plan E-commerce integrations on Pro and Enterprise

Instead of typing one word into a search box, the visitor describes what they need: "I'm looking for a jacket for city winters that isn't too bulky." The chatbot understands the request, finds suitable products in your catalog and recommends them with a short explanation of why they fit.

Recommendations come only from products in your catalog. Because Puhera cites the source of every answer, the visitor can see that the information comes from your product catalog.

• Google Merchant feed: The product feed you maintain for Google Merchant syncs your catalog to the chatbot. See the Google Merchant integration for details.
• T-Soft: The T-Soft integration includes catalog sync.
• Shopify and WooCommerce: E-commerce integrations connect your store to the bot for product recommendations, order lookup and similar actions.

Add content such as size charts and shipping and return policies to your knowledge base too, and the bot can handle the visitor's follow-up questions after a recommendation.

• "Which moisturizer do you recommend for dry skin?"
• "Do you have a gift for my dad under $50?"
• "Does this shoe come in a lighter color?"
• "What should I use together with this product?"

The bot answers from the price, color, variant and stock details in your catalog; it doesn't make up details your catalog doesn't contain.

Making recommendations is half the job; the other half is knowing whether they sell. Puhera tracks product view, add-to-cart, checkout and purchase events and uses SKU-based attribution to match each purchased product to a chatbot recommendation. A SKU is the stock code for each product or variant.

For the summary, funnel and top-products reports, see chatbot revenue tracking.

The bot knows what your catalog says, so the quality of its recommendations depends on the quality of your product data. These points make a difference:

• Descriptive titles: Instead of "Jacket 123", use titles that state the material, the use and the standout feature.
• Complete descriptions: Who it suits, how it's used and what it pairs with help the bot match needs-based questions.
• Separate variants: Give each color and size its own SKU.
• Current price and stock: Keep the feed or store data fresh so the bot doesn't suggest sold-out items.
• Consistent categories: Group similar products under the same categories and labels.

For help preparing your feed, read about building a chatbot product catalog from a Google Merchant feed.