Use case: e-commerce
AI product recommendations from a shopping assistant that asks first
Like a good shop assistant, the agent asks what the product is for, the budget and the size, then searches your live catalog and suggests a few matching products as cards with image, price and stock.
Visitor: I need a tent for two for windy coastal campsites. Budget around €300.
Agent: For windy coasts, our buying guide recommends a low tunnel or dome tent. In your budget I found the Fjell 2 at €279 and the Kyst 2 Pro, on sale at €299 instead of €349. Both are in stock.
Visitor: Which one is lighter?
Agent: The Fjell 2 weighs 2.1 kg and the Kyst 2 Pro 2.6 kg. The Kyst 2 Pro has a larger porch for keeping bags out of the rain, if that matters to you.
Visitor: I will take the Fjell 2. Do you deliver to Norway?
Agent: Yes, delivery to Norway takes 5 to 7 working days. You can order the Fjell 2 from its product page, and the customs details are on our shipping page.
Shoppers know what they need, not what it is called
A visitor wants a tent for two that copes with wind on the coast, but your filters ask for capacity, weight class and pole material. The search returns forty results or none. Many shoppers leave rather than compare product pages one by one.
Why add a shopping assistant to your store
It combines live product data with what you have written about your products.
Understands needs, not just keywords
Shoppers describe the use, the budget or the problem. The agent turns that into searches in your catalog.
Real products as cards
Suggestions appear as product cards with image, price, sale price, stock status and a link, built from what your store returned.
Advice from your own guides
The size charts, care instructions, materials and comparison guides you train the agent on shape its advice.
Your rules for recommending
Write in the instructions how to advise: ask for the budget first, suggest at most three products, mention products in stock first.
Advice at any hour
Shoppers get help in the evening and at weekends, when nobody is there to answer.
How to set up product recommendations
Connect your store
Paste your shop address. intoCHAT detects Shopify, WooCommerce, Magento, VTEX, WordPress or a custom site and adds live product search, with no API key or app.
Train it on your buying guides
Add size charts, materials, care and comparison pages, plus shipping and returns, so advice and policies come from your own content.
Write how it should advise
Start from the E-commerce Specialist template and add your rules: which questions to ask, how many products to show, what to say about sizes.
Test with real questions
Ask the playground what your customers ask, check the suggested products, then add the agent to your shop.
What powers the recommendations
Connect your store
Live product search on Shopify, WooCommerce, Magento, VTEX, WordPress and custom sites, from public catalog data.
Product details
Variants such as size or color, prices and stock, where the platform provides them.
Product cards
Up to 10 cards under a reply, with image, price, sale price, availability and link.
Trained knowledge
Buying guides, size charts, FAQs and policies from your website and files.
Custom API actions
Your own systems alongside the store connection, for example a delivery date lookup.
The shopper's language
The agent replies in the language the shopper writes in.
Frequently asked questions
How does an AI shopping assistant recommend products?
It asks about the shopper's needs, searches your live catalog with what it learned, and shows matching products as cards. It explains the choice using your product data and the buying guides you trained it on. The products shown always come from your store's search results.
Does it personalize recommendations based on purchase history?
No. It only knows the current conversation and your public catalog, not the shopper's account, past orders or customer-specific prices.
Can it add products to the cart or take orders?
No. The agent links to the product page, where the shopper adds the item to the cart and checks out as usual. It also cannot track orders or process returns, but it can answer questions about them from your policy pages.
Which shop platforms work?
Shopify, WooCommerce, Magento, VTEX and WordPress are detected automatically, and other sites can connect through their own search page or product listings. No API key or app install is needed.
Will it recommend products that are out of stock?
Product cards show whether an item is in stock or sold out when the store provides that information. You can also tell the agent in its instructions to suggest in-stock products first.
Can the shopper ask a person instead?
There is no live chat or human handoff. If a shopper needs personal advice, lead capture can collect their email or phone number, and you get an email alert so your team can follow up.
Help shoppers find the right product
Try the free store demo with your shop address, then create your agent free and connect your catalog.
Create your agent freeFree plan, no credit card needed.