💬Get free consultation

AI product recommendations

What are product recommendations

Product recommendations control which products your AI suggests during a chat. You can rely on the built-in sets, pull in Shopify's own recommendations, or write your own rules.

This is where you steer the AI toward the products you actually want to sell.

---

How to open product recommendations

1. Go to AI agent → Instructions → Assistant skills.
2. Under Shopping skills, find Product recommendations and click the edit icon.

Expected result: The page opens with System recommendations, the Shopify recommendations card, Custom recommendations, and Similar product matching.

---

System recommendations

These are built in and answer common shopping questions. You never pick the products yourself, but you do have to turn each one on.

Recommendation	Answers questions like
Best sellers	"What are your best sellers?"
New arrivals	"Any new items?"

Each row shows how many products it covers and when it last changed. Turn a row on with the Status toggle. Click the edit icon to see and change the products it uses.

---

How to turn on Shopify recommendations

Shopify already calculates product relationships from your store data. You can let your AI use them.

Type	What it suggests
Related products	Similar items customers may also like
Complementary products	Items that go well together

1. On the Shopify recommendations card, turn on the type you want.
2. Click the row, or the eye icon under Actions, to preview which products Shopify returns.
3. Click Save.

Expected result: Your AI can suggest related or complementary products during a conversation.

---

How to create a custom recommendation

Use a custom rule when you want specific products for a specific moment. For example, gifts, a seasonal sale, or a one-off occasion.

The card ships with two templates, Special occasions and Sale promotion. They already carry search terms, so editing one is faster than starting from scratch. A template is only saved once you edit it.

1. On the Custom recommendations card, click Add recommendation, or click the edit icon on a template.
2. In Title, name the rule. For example, "Valentine gifts under $50".
3. Under Search terms, add the phrases customers use. For example, "gift for her", "valentine present", "romantic gift".
4. Under Products conditions, pick the products or collections the AI should suggest.
5. Turn on Enable recommendation.
6. Click Save.

Expected result: The rule appears in the Custom recommendations table with its status on. The row shows how many products it covers and when you last changed it. When a customer mentions one of your search terms, the AI suggests those products.

To remove a rule you no longer need, click the trash icon on its row.

---

How to prioritize similar products by tag

By default the AI ranks similar products by automatic relevance. You can force it to favor products that share your tags.

1. Scroll to Similar product matching.
2. Turn on Prioritize similar by tags.
3. In Priority tags, pick the tags that matter. For example, "linen", "summer", "unisex". The counter next to the label shows how many you have used.
4. Click Save.

Expected result: When a customer asks for something similar to a product, the AI ranks items that share more of your tags first.

How many priority tags can I add?

Up to 50. Click Clear all to start over.

What if two custom recommendations match the same question?

Keep your search terms specific to each rule. Overlapping terms make the match unpredictable, so split them by intent instead of repeating the same phrase.

Do I need to add products manually for best sellers and new arrivals?

No. Both read from your synced catalog and update as your catalog changes.

My custom recommendation never fires. What should I check?

Check three things. The rule is set to Active, the search terms match how customers really phrase it, and the products you picked are in stock and synced.