What is Product Recommendation?
Product Recommendation is an AI feature that automatically suggests relevant products to your customers based on their needs. When you connect your shop, the AI accesses your product catalog, descriptions, pricing, and stock levels to make intelligent recommendations.
Every time a customer asks for help finding a product, your AI searches through your inventory and suggests 4-5 relevant items with direct links, images, and full descriptions.
Example: If a customer asks "I need a gift for my mom who loves skincare," the AI will analyze your product descriptions and recommend 4-5 skincare products that match, each with an add-to-cart link, product page link, description, and image.
How It Works
Automatic Product Access
When you connect your shop, the AI automatically:
Reads all product names and descriptions
Checks current pricing
Monitors stock levels
Only recommends in-stock products by default
Smart Recommendations
The AI uses your product descriptions to match customer needs with relevant items. The better your product descriptions, the more accurate the recommendations.
Each recommendation includes:
Product name
Direct add-to-cart link
Product page link on your website
Full description
Product image
No configuration required to start. You can activate Product Recommendation and let it run automatically. The AI handles everything without additional instructions. The optional settings below give you more control if you want it.
💡 To trigger product recommendations within a Process workflow, use the Recommend Product action — see article 2.4.
Setting Up Product Recommendation
Step 1: Activate the Feature
Go to Konvo AI → Recommendations
Toggle the feature ON
The AI immediately starts recommending products
📸 [Image: Screenshot of Recommendations page with toggle switch in green/ON position]
That's it—your AI is now recommending products automatically.
Customizing Product Recommendation (Optional)
If you want more control over how your AI recommends products, go to the Configure tab. All of these settings are optional.
Boost Products
Boosting prioritizes specific products or collections in recommendations. When the AI searches for products to suggest, it checks boosted items first.
When to use Boost:
Promote high-margin products
Push seasonal items
Highlight new arrivals
Clear excess inventory
How to add a Boost:
Click Add Another Boost
Choose Condition boosted: The criteria for boosting
Tags
Collection
Product name
Metafield
Material, color, or other attributes
Enter Condition value: The specific value to boost
Example: If Condition = "Tags", Value = "Premium"
All products with the "Premium" tag will be boosted
Adjust Boost level: The slider controls boost intensity
📸 [Image: Screenshot showing Boost products section with example boosts for "Premium" and "Archived" tags, including the three columns: Condition boosted, Condition value, and Boost level slider]
Example:
Condition boosted: Tags
Condition value: Premium
Boost level: +60
Now whenever a customer asks for product recommendations, items tagged "Premium" appear first if they match the customer's request.
You can add multiple boosts with different conditions.
Additional Information
This section lets you give the AI extra context about how to recommend products—similar to Custom Answers in Knowledge Hub.
When to use Additional Information:
Seasonal or themed recommendations
Specific selling personality or tone
Rules about product quantity per recommendation
Product pairing suggestions
How to add Additional Information:
Enter a Question / intent that triggers this information
Example: "Clarification if the user does not explain what they're looking for"
Write Instructions for the AI
Example: "If the user does not give you any indication about what they are looking to buy, first ask them some general questions about what kind of products they are looking for. We sell cosmetics, and fashion items from some of the merchants that already use Konvo. We also sell the coolest company merch ever, with konvo branded jumpers, t-shirts bags, water bottles, laptop sleeves and some other cool accessories."
Another Example:
Question / intent: "Personality and way of selling"
Instructions: "You are very charming and like to get to know your customers, but you're always trying to push for the sale. If the user shares something personal or some detail about why they are buying or who is it for you will make some joke or nice comment or compliment to validate what they said and thereafter you will try to convince them to buy. If they seem interested but not explicitly saying they want to buy, assume they do and proactively ask them if you should go ahead and create the cart for them."
📸 [Image: Screenshot showing Additional information section with "Personality and way of selling" and the corresponding instructions about being charming and pushing for sales]
You can add multiple Additional Information blocks for different scenarios.
Discount Codes
Add discount codes the AI can offer to customers in specific situations. This helps close sales by giving the AI incentive tools to use strategically.
⚠️ Critical: Discount codes must already exist and be active in your shop. The AI only shares codes—it cannot create them.
When to use Discount Codes:
Customer is hesitant about price
Customer is deciding between products
Encourage larger orders
Reward loyal customers
How to add Discount Codes:
Click Add Discount Code
Enter the Discount Code exactly as it exists in your shop
Example: FREE50
Write a Description explaining when the AI should offer this code
Example: "50% Discount on everything. When to use: Use when you realize the user is seems to want to buy something but is not decided because maybe pricing is a concern. Please do not use it otherwise."
Another Example:
Discount Code: SHIPPING100
Description: "Free Shipping in purchases with an order value > 50€. When to use: Offer it when the user wants to buy something that is worth less than 50€ but close to it, so that we can incentivize them to order more and if it crosses over 50€ then we can offer free shipping"
The AI will evaluate each conversation and decide when offering a discount makes sense based on your instructions.
You can add multiple discount codes with different use cases.
Best Practices
Starting with Product Recommendation
Step 1: Just activate it
Turn on Product Recommendation and let it run for a few days. The AI performs well with zero configuration.
Step 2: Monitor performance
Review conversations where the AI recommended products. Look for:
Are the recommendations relevant?
Is the AI recommending the products you want to highlight?
Are customers responding positively?
Step 3: Add boosts strategically
If certain products aren't appearing enough in recommendations, boost them.
Step 4: Refine with Additional Information
Add personality or specific recommendation rules if needed.
Step 5: Test discount codes carefully
Start with one discount code and monitor how often the AI uses it. Add more once you're confident.
Writing Instructions for Additional Information
✅ Do:
Use "Explain to the customer that..."
Use "Recommend products that..."
Provide clear, specific guidance
Include examples of scenarios
❌ Don't:
Use "Say: '[exact text]'"
Force the AI to copy exact quotes
Overload with unnecessary instructions
Example:
❌ Bad instruction: "Say: 'Our Christmas collection features beautiful red and green items perfect for the holidays.'"
✅ Good instruction: "When customers mention Christmas, holidays, or gift-giving in December, prioritize recommending products in red and green colors. Mention that these items fit the festive season."
Discount Code Strategy
Conservative approach (recommended):
Add 1-2 discount codes
Use them for clear scenarios (hesitant buyers, cart value thresholds)
Monitor how often the AI offers them
Aggressive approach:
Add multiple discount codes for different situations
Give the AI flexibility to close more sales
Accept higher discount usage
⚠️ Important: The AI will follow your instructions. If you write "offer this discount when the customer seems interested," the AI will use it frequently. Be specific about when discounts should be offered.
Boost Strategy
Don't over-boost:
Boosting too many products dilutes the effect. Focus on 2-4 key boosts maximum.
Strategic boosting:
High-margin products
Seasonal items (boost in season, remove after)
New launches
Slow-moving inventory you want to clear
Example seasonal approach:
November-December: Boost "Holiday" collection
January: Remove holiday boost, add "New Year" collection
February: Boost Valentine's products
How Your AI Uses Product Recommendation
Semantic Understanding
Your AI doesn't just match keywords—it understands customer intent.
Example:
Customer: "I need something warm for winter"
AI understands: Customer needs winter clothing/accessories
AI searches: Products with descriptions mentioning "warm," "winter," "cold weather," "insulated," etc.
AI recommends: 4-5 relevant items from your catalog
The AI reads your product descriptions to understand what each item is for, then matches those to customer needs.
Recommendation Trigger
The AI only recommends products when:
The customer expresses interest in buying something
The customer asks about specific product types
The conversation context suggests product recommendations would help
The AI won't spam every message with products—it waits for the right moment.
Stock Awareness
By default, the AI only recommends in-stock products. If a product is out of stock, the AI skips it and suggests alternatives.
Troubleshooting
AI isn't recommending the right products
→ Check your product descriptions—are they detailed and accurate?
→ Add an Additional Information block with specific guidance
→ Use Boost to prioritize certain products
AI recommends out-of-stock items
→ This shouldn't happen by default, but verify your shop's stock levels are syncing correctly with Konvo
AI offers discounts too often
→ Review your discount code descriptions and make the "When to use" instructions more restrictive
AI doesn't recommend products at all
→ Verify Product Recommendation is toggled ON
→ Check that your shop connection is active and products are syncing
Advanced Tips
Pairing Boost with Additional Information
Combine these features for powerful control:
Boost your "Summer Collection"
Add Additional Information: "When customers mention beach, vacation, or hot weather, emphasize products from the Summer Collection."
This ensures summer items appear first AND the AI talks about them in context.
Seasonal Workflow
Create a calendar for updating Product Recommendation:
January: Boost New Year / Fresh Start products
February: Boost Valentine's gifts
March-April: Boost Spring collection
May: Boost Mother's Day gifts
November-December: Boost Holiday / Christmas collection
Toggle boosts ON and OFF as seasons change rather than deleting them.
Testing New Products
When launching new products:
Add a boost for the new product tag/collection
Add Additional Information: "We just launched [product name]. When relevant to customer needs, mention it's new and recommend it enthusiastically."
Monitor conversations for 1-2 weeks
Adjust boost level based on performance
What to Explore Next
→ 2.6 Testing your AI