Use case · Growth · Sales
Dynamic Discount Offers
Dynamic discount offers give one shopper a code for one product, at the moment they're close to buying. An AI chat agent spots the interest, a tool creates a unique code in your store through its API, limited to that product, one use per customer and a short window, and the agent offers it in the conversation.
- Each code works for one product only
- 1 product
- Per customer, then it's spent
- Once
- Code lifetime in our Shopify demo
- 3 days
What are dynamic discount codes?
Dynamic discount codes are unique, single-use codes created on demand for one shopper and one product. Instead of a site-wide code that ends up on coupon sites, an AI agent generates a code when a shopper shows clear interest, your store creates the matching discount through its API, and the code expires after a short window.
Why do site-wide discount codes leak?
Site-wide discount codes leak because anyone can share them. They end up on coupon sites, get used by customers who would have bought anyway and teach shoppers to wait for the next sale. Yet a shopper weighing up one product is often a small nudge from buying, and nobody can create one-off codes by hand at scale.
Codes end up on coupon sites
A code meant for a few people spreads to everyone.Discounts go to people who'd have paid
Margin is given away on sales you'd have made anyway.Undecided shoppers get nothing
The one shopper a nudge would help sees the same page as everyone else.
How does an AI agent create a one-off discount code?
An AI chat agent creates a one-off code in four steps. It spots a buying signal after recommending products, takes that product's ID and name from the conversation, has a tool generate a random code with a start and end time, and an automation creates the discount in your store. The agent then shares the code, value and expiry.
With an AI chat agent
A shopper shows clear interest in a product the agent has recommended
Spot the interest
Buying signals such as choosing a product, asking how to order it or comparing it closely. When an offer is made is set in the agent's instructions.Pick the right product
The product's ID and name come from the products the agent has already shown in the conversation.Create the code
A tool generates a random code with a start and end time, then an automation creates the discount in your store through its API: that product only, once per customer, for a short window such as 3 days.Offer it in the chat
The agent shares the code, the discount value, the product it applies to and when it expires, so the shopper can buy straight away.
With AI automations
A set event, such as an abandoned cart, a first purchase or a win-back date
Call the same discount tool
Other workflows can use the discount automation too.Create a unique code
For the product left in the cart, or for the customer's next order, created in your store.Add it to the message
The code goes into the email or SMS that goes out.
What rules can you set?
You set every rule the agent works within: the discount type, a percentage or a fixed amount, the value, how long codes last, whether each customer can use a code only once, and which products or shoppers qualify. The agent only offers discounts inside those limits, and changing the rules means editing one place.
- Type and value. A percentage or a fixed amount, and how much.
- Lifetime. How long each code lasts, for example 3 days.
- Usage. One product, once per customer.
- Eligibility. Which products qualify, which shoppers, and how often one shopper can be offered a code.
- Maximum value. A ceiling the agent can't go above.
Which systems deliver it?
Dynamic discount offers are delivered by an AI chat agent, which decides when to offer a code and shares it, and by AI automations, which create the discount in your store through its API. The same discount automation can be called from email and SMS workflows too, such as abandoned carts and win-back campaigns.
| System | Trigger | What it does | Package note |
|---|---|---|---|
| AI Chat Agents | Shopper shows clear interest | Spots the signal, picks the product, offers the code with its value and expiry | Builds on a recommendation agent with live catalogue look-up, an add-on on any package |
| AI Automations | Code request from the agent, or a set event | Creates a single-use, product-specific, time-limited discount and returns it | Needs API access to your store's discounts |
Where is it used?
Dynamic discount offers suit online stores and D2C brands on Shopify or another platform with a discount API, especially those selling higher-value products where a small, targeted nudge closes the sale. E-commerce teams call them promo codes, discount codes or personalised offers. The difference here is that each code is made for one shopper.
Same job, their words
Personalised offers in the store chat. A shopper compares two jackets, picks one and asks about delivery; the agent offers a code for that jacket only, valid for a few days, and the shopper checks out in the same visit.
How would we build it for you?
We built this as a demo for a Shopify store, alongside a product recommendation agent, in a test project for a Make.com community challenge. We haven't built it for a client yet. A build for you starts by agreeing the offer rules, then connecting your store's discount API and testing codes before anything goes live.
- Shopper picks a productbuying signal
- Start time setUTC · ends in 3 days
- Random code made8 characters
- Discount createdShopify GraphQL
- Code returnedvalue · expiry
- Offered in chatready to buy
In the demo, a custom function in Voiceflow found the current time in UTC, set the code to expire 3 days later and generated a random 8-character code. A Make.com scenario created the discount with a single Shopify GraphQL mutation, a percentage off the chosen product, usable once per customer, so changing the rules means editing that one module.
What should you know before you start?
Dynamic discount codes need two things before anything is built: rules you've agreed and API access to your store. The agent should only offer discounts within limits you set, such as which products qualify, how often a shopper can be offered one and the maximum value. Every code is created in your store and tracked like any other discount.
- Agree the rules first. Products, frequency and maximum value, written into the agent's instructions.
- Platform access. Your store needs a discount API, as Shopify has, and the automation needs access to it.
- Tracked like any discount. Codes sit in your store's own discount list, so you can see which were used.
Frequently asked questions
Questions
Answer
Can a dynamic discount code be shared or reused?
Not usefully. Each dynamic discount code is created for one product, can be used once per customer and expires after a short window, such as 3 days. If it's posted on a coupon site, it only takes money off that one product, for a short time, so it's worth far less to anyone else than a site-wide code.
Will the agent offer a discount to everyone?
No. The agent only offers a code when a shopper shows the buying signals set in its instructions, such as choosing a product or asking how to order it, and only within the limits you agree: which products qualify, how often one shopper can be offered a code and the maximum value. Shoppers who are ready to buy can buy at full price.
Does it work with platforms other than Shopify?
It works with any store platform that lets an automation create discounts through an API. Our demo was built on Shopify, using its GraphQL API to create a discount for one product, once per customer. For another platform, we check its discount API on the discovery call before anything is scoped.
Can codes go into abandoned cart emails?
Yes. The same discount automation can be called from other workflows, so an abandoned cart, first purchase or win-back date can trigger a unique code for that product or the customer's next order. The code is created in your store and added to the email or SMS. See email marketing automation for the flows themselves.
Have you built this for a client?
Not yet. We built it as a demo for a Shopify store, in a test project for a Make.com community challenge, alongside a product recommendation agent and order tracking. The demo created real discounts in the store through its API. A client build follows the same design, with your rules and your platform.