Gorgias logo
Gorgias logo

All articles

How AI Agent recommends products and offers discountsUpdated 44 minutes ago

When a shopper visits your website, AI Agent acts like a smart, personalized shopping assistant that’s available to answer questions and help them find and buy products from your store.

During its conversations with shoppers, AI Agent can make tailored product recommendations and offer discount codes to guide them toward a purchase.

In this article, you’ll learn how AI Agent makes decisions to recommend products, give discounts, and how it adapts in real-time to carry out sales conversations.

Requirements

  • You must have an AI Agent subscription

    • If you subscribed to AI Agent prior to July 16, 2025, you may need to upgrade your subscription to access shopping assistant features
  • You must have a Shopify store connected to Gorgias

Context-based decision making

When AI Agent makes a product recommendation or offers a discount code, it doesn’t do so at random. Like a good sales associate, it understands that recommendations and discounts are most effective when they are responsive to the shopper’s needs and well-timed.

To do this, AI Agent observes the shopper’s browsing activity to build an adaptive profile. This is what we mean when we say AI Agent context aware. It dynamically learns who the shopper is, what they’re interested in, and watches for signals to identify their level of interest or buying intent.

AI Agent can then make real-time decisions during a conversation, including knowing when to recommend products and use discounts.

Shopper browsing activity

To tailor its responses and recommendations, AI Agent builds a profile of the shopper using browsing activity from their current session on your website. Browsing activity gives AI Agent signals about the shopper’s interest and motivations — is the shopper just browsing, looking for something specific, or ready to buy?

Behavioral signals that AI Agent looks for, include:

  • Products viewed → which product pages have they visited?
  • Pages views → what pages have they explored, like FAQs, reviews or categories?
  • Current page → which page are they currently browsing?
  • Purchase history → what have they bought from this brand before?

    • Note: the shopper must be signed in to your Shopify store for AI Agent to access their purchase history
  • Cart data → what items are currently in their cart?

Levels of buying intent

Using their browsing activity, AI Agent is able to determine the shopper’s buying intent.

A shopper’s buying intent allows AI Agent to adapt the conversation in real-time. AI Agent can understand when it needs to ask more questions, when to educate, and when it's appropriate to make recommendations or offer a discount code.

The buying intent of a shopper can dynamically change based on their conversation with AI Agent and their ongoing activity on your website. There are three levels of buying intent:

StageKey behaviors
Discovery

The shopper is browsing. No clear signal of sales opportunity

  • Vague questions
  • Broad exploration of the store
  • Questions about logistics (delivery, refunds) with an empty cart
  • Restock inquiries without naming specific products
Interested

The shopper has a clear interest in a product but hasn’t taken any purchasing action

  • Mentions specific product or category
  • Questions about differences, compatibility, variants and restocks
  • Questions about sizing and returns while looking at specific products
Ready to buy

The shopper has strong purchase intent

  • Cart contains products
  • Questions about checkout, shipping and discounts
  • Expressing a clear intent to buy, even without items in cart — for example “I want to buy this dress in size small”
  • Waiting for restock to place an order

Your selling style preferences

Your selling style settings are a significant factor that directly affect how AI Agent interacts with shoppers and adapts its conversations. You can choose an Educational, Moderate or Promotional style to match your brand’s approach to selling.

For example, here’s how AI Agent would interact with a shopper with a Discovery buying intent, where your selling style is set to Educational:

  • Ask refining questions: AI Agent will focus on learning about the shopper’s needs, sharing product details and comparisons
  • Recommend products only when confident: AI Agent may only make a specific product recommendation when the shopper’s needs are clear
  • Avoid discount codes: AI Agent won’t use discounts at this stage to incentivize a sale

Review the following table to see how conversations might change based on the shopper’s intent and your settings for AI Agent:


How selling style affects different level of buying intent

Discovery (casual)Interested (considering)Ready to buy (high intent)
EducationalAI asks open-ended questions and seeks to understand shopper needsAI provides detailed product information and comparisons to help shoppers decideAI offers support with checkout, but no urgency
Moderate AI asks a mix of questions and makes light product recommendationsAI suggests products and may highlight promotions or bestsellersAI encourages purchase completion with a friendly nudge, such as limited-time discounts
Promotional AI quickly suggests products based on minimal input, focusing on conversionAI makes deliberate product recommendations and emphasizes urgencyAI uses strong calls to action to create urgency and close the sale

Making product recommendations

AI Agent recommends products from your store based on signals from your shopper’s browsing activity and their buying intent. Products that AI Agent recommends may include:

  • Similar products to something the shopper mentioned — “Do you have something like the Nimbus 25?”
  • Complementary items that go well with what the shopper is currently browsing — “example sentence”
  • Popular or trending products when the shopper is just browsing or the AI doesn’t find products that closely match what they’re looking for — “example sentence

Recommendations on Chat and Email

AI Agent shares recommendations differently based on the channel where the conversation with the shopper is happening.

  • Conversations on Chat: AI Agent shares a visual Shopify product card. Shoppers can select the card to go to the product page or add the item directly to their cart.
  • Conversations on Email: AI Agent shares a link to the recommended product’s page. It does not include product images.

AI Agent message in a chat window recommending Haabitual cream for dry skin with an image of a blue product tube, price listed as $99.00, and an “Add To Cart” button.

Products excluded from recommendation

Some products are automatically excluded from recommendations. These include:

Prevent AI Agent from recommending specific products

If you have a product(s) that you do not want Shopping Assistant to recommend, you can exclude the product using a tag in Shopify.

Go to your products in Shopify. Under “product organization” add gorgias_do_not_recommend as a tag to the product. Shopping Assistant will no longer recommend the product in conversations so long as the tag remains.


Sharing discount codes

AI Agent uses what it knows about your shopper to decide if and when to offer a discount. It uses the shopper’s buying intent, plus signals from their browsing activity to decide when a discount could help lead to a sale.

Buying signals that AI Agent looks for include:

  • What page you’re on (e.g. checkout vs. product browsing)
  • Whether there are items in your cart
  • If you’ve shown strong buying signals (like saying “I want to buy” or asking about shipping)
  • If you’ve hesitated after showing interest

Discount frequency and amount

Your discount strategy settings are a significant factor that directly affects how frequency AI Agent shares discount codes. The discount code amount is a fixed percentage (%) that you set in your AI Agent’s settings.

You can choose one of four strategy settings for sharing discounts. Each strategy controls how often an AI Agent can offer a discount, and under what conditions.

The following table explains what you can expect from each strategy level:

Strategy typeDescription
No DiscountNever offer discounts
ConservativeOffer discounts only in rare cases, like when a high-intent shopper hesitates at checkout
StrategicUse discounts selectively based on shopper behavior and likelihood to convert. Prioritize timing of discounts, like in key moments of hesitation or price comparison
GenerousUse discounts often to maximize conversions and reduce cart abandonment. Any hesitation may trigger a discount.

How discount codes are created

If AI Agent decides to use a discount code to incentivize a sale, it creates a new unique discount code in your Shopify admin.

  • The value of the discount code is based on the fixed amount that you set up in your discount strategy settings
  • Each discount code is valid for 3 days and limited to a single use
  • AI Agent shares the code with the shopper in their conversation. For example: “if you’re ready to check out, here’s a 10% discount to help you out: SAVE10-TGQ9
  • You can review all of the discount codes that AI Agent shares with shoppers and their usage in your Shopping Assistant analytics

Chat interface where the AI Agent offers an 8% discount on a product called “Shimmer Layer,” with a visible coupon code and a price of $52.00, alongside a settings panel to configure discount tone and maximum percentage.


Was this article helpful?
Yes
No