A blog by StabileRewards.

Blog

Your Shopper’s AI Can Read a Price. Can It Explain the Reward?

Cover Image for Your Shopper’s AI Can Read a Price. Can It Explain the Reward?
Nova Mason
Author
Nova Mason
Mina Gemelli
Editor
Mina Gemelli

Direct Answer An AI shopping assistant should explain a loyalty offer’s earning rate, minimum spend, conditional bonuses and timing. Points an order may earn, a reward already issued and a discount accepted at checkout are different kinds of value. Stabile Rewards Wallet already provides public reward terms and cart-earning forecasts for participating stores. We’re exploring clearer first-order bonus explanations and minimum-spend gaps, and gathering merchant feedback before building further.

A shopper asks an AI assistant:

“What would this $72 order earn at your store?”

“510 points” sounds precise. But perhaps 360 points come from the purchase itself, while another 150 are available only on an eligible first order.

The shopper needs to know which conditions apply and when the benefit can be used. Otherwise, an accurate calculation can still leave them with the wrong expectation.

For a merchant, that is a familiar explanation problem in a new setting. The earning rules you describe on a rewards page also need to make sense when an assistant answers a customer’s question.

What should an AI assistant explain about loyalty points?

Consider this illustrative USD loyalty program, using synthetic numbers:

Program ruleIn this example
Earning rate5 points per $1 of qualifying subtotal
Minimum to earn$35 qualifying subtotal
First-order bonus150 extra points on an eligible first order
Reward threshold500 points qualify for a $5 reward on a future order, subject to the program’s rules

For a qualifying $72 order, base earning is 360 points. With the first-order bonus, the projected total reaches 510 points—enough to cross the example program’s reward threshold. Without the bonus, the order earns 360 points.

A $28 cart is $7 below the earning minimum. Under this example, it earns no points. Reaching $35 would make the subtotal eligible to earn; even with the first-order bonus, that order would earn 325 points, below the 500-point reward threshold.

A more useful answer to the original question would be:

“This qualifying $72 order would earn 360 points. If it’s your first eligible order, you’d earn another 150, bringing the total to 510—enough to reach this program’s $5 reward threshold for a future purchase. Your first-order eligibility hasn’t been checked.”

That last sentence matters. A public program lookup can explain a bonus, but it cannot establish a particular shopper’s purchase history. A statement such as “I’m new here” should not be treated as verified eligibility.

The answer gives the customer something useful: an estimate, a clear condition and an explanation of when the reward could help.

How do projected points differ from an available reward?

Reward information can describe an opportunity, an existing benefit or an accepted discount. Each answers a different shopper question.

Benefit typeShopper’s questionWhat the answer should explain
Projected earning“What would this order earn?”Earning rate, minimum spend, exclusions and conditional bonuses
Reward already issued“What rewards do I have?”Availability, expiry and redemption conditions
Discount accepted at checkout“Did my reward work?”Accepted discount amount and the checkout’s returned totals

A forecast does not credit points. An issued reward still needs to meet the store’s redemption conditions. Checkout confirms whether a discount applies to this purchase.

Keeping these answers separate helps an assistant explain future value without presenting it as money already available to spend. It also makes the next step clearer: check eligibility, retrieve an existing reward, or review checkout’s response.

Where does UCP fit into reward explanations?

Shopping systems need a consistent way to exchange offer information. Universal Commerce Protocol (UCP) provides a shared framework for communication between commerce systems and AI agents.

One proposal, the UCP incentives extension, explores describing offers during shopping, including offers a cart is approaching and the remaining gap. As of November 15, 2026, it is a draft.

That direction is relevant to our question: can an assistant explain an offer’s conditions while a shopper considers a purchase? Future points and discounts on the current purchase are different benefits, so an earning forecast cannot simply be placed into a field intended for an immediate discount.

Availability also depends on support from the commerce platform and shopping agent. Shopify’s negotiation model selects capabilities both sides support; an agent’s declaration alone cannot enable a feature at a store.

We’re exploring how these ideas could fit Stabile, with the merchant’s rules and checkout remaining authoritative.

What can Stabile Rewards Wallet do today?

Stabile Rewards Wallet can show a participating store’s public reward terms and estimate what a cart would earn without shopper sign-in.

Personal saved rewards and codes use a separate sign-in and store-linking flow. A shopper can ask about rewards from linked stores, then use a code at the store’s normal checkout. Our Wallet launch article explains that flow and its limits.

These existing capabilities give us a starting point. The next question is which explanation would be useful enough for merchants to put in front of customers.

What improvements are we considering?

We want to test two ideas:

  1. Clearer first-order bonus explanations. Show base earning separately from a bonus whose eligibility has not been checked. If the bonus takes the order past a reward threshold, carry that condition into the explanation of the future reward.
  2. A clearer gap to the earning minimum. Explain how much a cart falls short, using the store’s currency and qualifying-subtotal rules. Distinguish the minimum to start earning from the points needed to reach a reward.

These are proposed improvements. We’re also evaluating a synthetic mapping to the emerging UCP incentives model; that work does not establish a live Shopify integration.

FAQ

Does a reward forecast reduce the current checkout total?

No. A forecast estimates points a qualifying order may earn. A future reward and a discount accepted on the current purchase are separate benefits.

Can a shopper see public Stabile reward terms without signing in?

Yes, for participating stores that make their program available to the assistant. Personal saved rewards and codes require the separate sign-in and store-linking flow.

Does reaching the earning minimum guarantee a reward?

No. The earning minimum determines when an order starts earning points. The program’s reward threshold determines when enough points qualify for a reward. In this article’s example, those are $35 of qualifying subtotal and 500 points, respectively.

What do your shoppers misunderstand?

Do customers confuse points they may earn with rewards they can use now? Do they miss the earning minimum? Does a first-order bonus cause the most questions? Or does the confusion happen later, when they try to use a code?

We’re gathering store and program examples to decide what to evaluate next.

Would a clearer explanation of what an order earns help your shoppers? Tell us your store and the reward condition customers most often misunderstand. Mention “reward clarity” through the Stabile contact page.

Bring your store’s rewards into AI-assisted shopping

If you want to try Stabile’s existing reward terms and Wallet features, install Stabile’s Shopify app, then message us in the app about AI reward visibility and Wallet setup for your store.

Already using Stabile? Message us in the app and we’ll help you get set up. The clearer first-order bonus explanations and minimum-spend gaps discussed above are proposed improvements to those existing features.

Sources

Share