Skip to main content
Docs/AI Recommendations
DocsAI recommendations
Free

AI recommendations

Show shoppers products they're likely to buy: 'similar', 'frequently bought together', and personalized picks.

Recommendations surface the right products at the right moment: related items on a product page, a "frequently bought together" bundle in the cart, and personalized picks for a returning shopper. It is the "you may also like" experience that nudges a browser toward checkout.

The fastest way to add it is to ask in chat, for example "add ‘you may also like’ suggestions on my product pages and in the cart." GenMB wires the suggestions into your app for you, and you can fine-tune the wording and placement from there.

Recommendations build on your product catalog and shopper activity. Add your products with the Ecommerce SDKs first, then let the app record what people view, add to cart, and buy so the picks keep improving.
Product pages and the cart are where recommendations lift sales the most. Until there is enough activity to learn from, the app shows popular products so the section is never empty.

For developers

The suggestions are available through window.genmb.recommend. Record activity with trackEvent so the picks improve over time.

// Record a shopper action (view | add_to_cart | purchase)
await window.genmb.recommend.trackEvent({
  userId: 'u_123',
  type: 'view',
  productId: 'product_42',
});

Then fetch suggestions where you want to show them:

// Related products on a product page
const { data } = await window.genmb.recommend.similar('product_42', 4);

// Personalized picks for a returning shopper
const { data: forYou } = await window.genmb.recommend.personalized('u_123', 5);

FAQs

What do I need before recommendations work?
A product catalog and some shopper activity. Recommendations are built on top of the Products SDK, so add your products first. The suggestions get better as your app records what people view, add to cart, and buy.
What if there is not enough data yet?
When personalized results are not available, the SDK falls back to popular products so a "you may also like" section never shows up empty. As more activity comes in, the suggestions become more tailored.