Do Shoppers Ask Chatbots for Discounts? Yes, More Than Anything Else (2026 Data)

TL;DR: On live online stores, "do you have a discount code?" is the most common thing shoppers type into a chat assistant: 7.0% of all visitor messages, more than product recommendations (4.7%), price (4.6%) or "where is my order" (4.2%). The data comes from 9,480 real visitor messages on Chatonbo storefront assistants over the 60 days to September 19, 2026. A store that treats the discount question as a nuisance loses a shopper who was one sentence from buying. A store that answers it well, with a real code or an honest "no code, but here is the best value", turns it into a checkout.

What shoppers actually type into a store chat

We looked at every visitor message on live storefront assistants for 60 days and classified the intent of each one. Most messages are ordinary product questions, greetings and follow-ups. Among the transactional intents, this is the order:

• Discount, coupon or promo code: 7.0% of messages
• Product recommendation ("which one should I get"): 4.7%
• Price: 4.6%
• Order status ("where is my order"): 4.2%
• Shipping: 2.7%
• Sizing: 2.6%
• Talk to a human: 1.3%
• Stock: 0.8%
• Complaints: 0.4%

The discount question wins on every kind of store we looked at, from massage chairs to apparel to supplements. It is not a sign of a bargain-hunting audience. It is what people do when they are close to buying and want a reason to finish.

How badly it goes wrong

We also measured how often the assistant failed to give a usable answer to each intent. Discounts failed 11% of the time, order status 16%, stock 21%. The discount failures were almost all the same shape: the store had no code configured, and the assistant either apologised at length or, worse, tried to be helpful and made one up. A made-up code fails at checkout, and that shopper does not come back to the chat to ask why.

Since September 19, 2026 every Chatonbo assistant on a store without codes answers in one sentence that there is no active code right now, never guesses one, and offers to email the shopper when a promotion starts. That single rule removes most of the 11%.

What a good answer looks like

There are three good answers, in order of how often they apply:

1. A real code, validated and applied. When the store has a code, the assistant checks it against the store before offering it, and on Shopify it can add the product to the cart with the code applied. One code per conversation, offered when the shopper hesitates or asks, not in the greeting.
2. No code, but the best value. "There is no active code right now. If you want the best value today, the 4-pack is $6 per unit instead of $8." Honest, short, and it keeps the shopper on the product.
3. No code, but a reason to stay in touch. "No code right now. If you leave your email I will let you know when the next promotion starts." That is a lead, and 12.8% of all conversations end with contact details for exactly this kind of reason.

What we changed in the product because of this

Three things, all live: a rule that forbids invented codes on stores without one; a validate-before-offer step for stores with codes; and on product, cart and collection pages, one-tap questions under the greeting so the discount question and the shipping question are a single click. The full set of figures, with data sets and dates, is on AI chatbot statistics 2026, and the wider study of what shoppers ask is in the full study.

What to do on your store this week

• Decide whether you want a code in the chat at all. If yes, one small code, validated, applied in the cart.
• If no, make sure the assistant says so in one sentence and offers the best-value product or an email alert.
• Read ten real discount conversations from your own chat. If the assistant is apologising in three paragraphs, shorten its instructions.