AI Recommendations
The right products, in the places that convert.
Try it, step by step
Where the right products appear
Illustrative store and products. The screens, wording and behaviour are the module's own.
What it does
Frequently Bought Together and Similar Products on the product page, You May Also Like on home and category, Customers Also Bought in the cart, and Post-Purchase Picks on the order-success page.
Learns from what your shoppers actually do — views, carts, purchases — rather than from a fixed rule someone configured once.
Personalised blocks inside transactional and newsletter emails, with the revenue they drive attributed back.
Measured against a control group that gets no AI, so the lift you see is the lift you got.
How it works
-
1
Shopper behaviour is recorded anonymously
Product views, add-to-carts and purchases, tied to an anonymous visitor ID. No names, no emails.
-
2
Patterns are learned per store
What is bought together, what is viewed together, what similar shoppers chose. Your catalogue, your customers — not a generic rule.
-
3
Blocks render where they convert
Frequently bought together and similar products on the product page, customers also bought in the cart, picks after the order.
-
4
The lift is measured
A small share of visitors see no AI. The Magento dashboard badge shows the revenue difference, with a confidence level, not a guess.
What you need
- Magento 2.4.4 or later with cron running
- Or Medusa v2, Shopify or WooCommerce 7.0+, with the Devexa connector
- A few weeks of traffic for personalised results (best-sellers are used until then)
- A Devexa plan that includes Recommendations (Business or higher)
Set it up
- 1 Install the modules and add your licence key.
- 2 Under Stores › Configuration › Devexa › AI Recommendations, choose which blocks appear on which pages.
- 3 Optional: add the email block to your order and newsletter templates.
- 4 Check the ROI badge on your Magento dashboard after a few weeks.
Screenshots and recordings
Questions merchants ask
How long before recommendations are good?
From day one they use best-sellers and similarity. Personal patterns appear as your store collects views and orders — usually within a few weeks.
How do I know it actually sells more?
A control group sees no AI. The dashboard compares revenue per visitor between the two and says when the difference is statistically significant.
Does it work with full-page cache?
Yes. Shared blocks are cached; personalised blocks for logged-in customers bypass the cache as Magento does for them anyway.
Which plan includes it
Included in: Business, Professional, Scale
From €103 per month, billed annually.
Ready to Grow Your Store?
Works with your store
One platform, whichever store system you run
The same search, shopping assistant, recommendations, visualiser, banners and form protection on every store system. Your catalogue syncs on its own and the widgets take your theme's colours.
Magento 2 / Adobe Commerce
The Devexa Connector for 2.4.4 to 2.4.9, Hyvä and Luma: install it with Composer, paste the key, and configure the widgets here on the platform.
Medusa v2
One npm plugin. Sales channels become stores, the settings are on the platform, the widgets drop into any Next.js storefront.
Shopify
An app from the Shopify App Store: paste the licence key and the widgets appear in your theme.
WordPress / WooCommerce
A plugin under Settings → Devexa in wp-admin, for WooCommerce shops.
Choosing a platform?
Need something modern for e-commerce? Build it on Medusa.
Medusa is an open-source commerce engine with a fast headless storefront, no licence fees and no plugin lock-in. Devexa is fully integrated with it: everything we offer works on a Medusa store from day one.
Talk to us about Medusa- Catalogue, stock and orders sync on their own
- Search, SmartAssist, recommendations and the Visualizer built in
- Promo banners and form protection, configured on the platform
- Your own storefront and checkout, nothing rented