Your price was written for a person scrolling a product page. More and more often, the first thing to read it is an AI assistant comparing you with every other store at once.
That shift is already measurable. Traffic to US retail sites from generative AI tools rose 693.4% year on year over the 2025 holiday season, according to Adobe. The base is still small, but the direction is clear. And since OpenAI launched Instant Checkout in September 2025, a shopper can go from question to purchase without leaving ChatGPT.
So it is worth asking a simple question before the next peak season: when the shopper is an assistant, what happens to a fixed price?
The assistant shops differently from a person
A person browses a few tabs, gets distracted and often buys from the store they trust. An assistant does the legwork in seconds. It reads the request (“a refurbished business laptop under £400 with 16GB of RAM”), searches widely and returns a shortlist.
Three things change for merchants:
- Comparison becomes total. The shopper no longer needs to open ten tabs. The assistant has effectively opened all of them.
- Price becomes a filter, not a story. “Under £400” is a hard line. A product at £415 may not be shown at all, however good it is.
- The product page loses some of its power. Your photography, reviews and copy still matter, but the first judgement is often made on structured data: price, availability, specification.
OpenAI says ChatGPT’s product results are ranked on relevance, not paid placement. When several merchants sell the same item, it weighs price alongside availability, quality and whether the merchant is the primary seller. Price is not the only factor, but in a side-by-side comparison it is one of the easiest to read.
Why a fixed price is a blunt signal
A fixed price says one thing: this is the number, take it or leave it. To a person, it sits inside a wider experience of brand, reviews and trust. To an assistant working through a shortlist, it is mostly a number to compare.
That creates two problems.
First, you cannot respond to the context. The shopper may be buying three units, may be a returning customer, or may be happy to take an older model. A fixed price cannot see any of that, so it treats every shopper the same.
Second, the usual fix is a blunt one. When conversion dips, most stores reach for sitewide discounts or voucher codes. Discounts are already deep at peak: Adobe recorded peak discounts of 30.9% on electronics and 25.1% on apparel in the 2025 holiday season. A discount that every shopper and every assistant can see quickly becomes the new expected price.
None of this means fixed prices are wrong. For most of a catalogue they are exactly right. The question is whether every product needs to work the same way.
Price as a range, not a single number
The alternative is to treat price, for some products, as a range you control. You set the list price and the lowest price you would accept. Anything in between is open to a conversation, with a person or with the assistant acting for them.
This is not new. Car dealers, B2B sellers and marketplaces like eBay have worked this way for years. What is new is that the conversation can now happen instantly and at scale, rather than over email days later.
In practice, merchants who do this well make a few decisions up front:
- Which products are open to offers. High-ticket items, slow-moving stock, refurbished goods and one-off pieces usually suit it. New-season bestsellers usually do not.
- Where the floor sits. Base it on margin, not gut feel. The floor is the number you would be happy to see on every order.
- What changes the floor? Stock levels, time on shelf, quantity bought and season are all sensible inputs. A product that has sat for 90 days can reasonably flex more than one that arrived last week.
- What happens below the floor. Rather than a flat no, the offer can be passed to a person to approve or decline. You keep the final say on the cases that need judgement.
One reseller I spoke to shows how this works in real life. He sells on eBay and uses an AI assistant to sift through incoming offers, then accepts or declines each one by hand. His rule is simple: when he is behind his monthly target, he accepts more offers; when he is ahead, he holds firm. In other words, his floor already moves with his budget. He just applies it manually, one message at a time.
Done this way, the shopper gets an answer in the moment, and the merchant never sells below a number they chose.
The risks, and the guardrails that matter
Flexible pricing is not free of risk. Four come up most often.
- A race to the bottom. If assistants learn that every price bends, they will always ask. The answer is a firm floor and a clear list of products that are not open to offers.
- Damage to brand value. A luxury or premium brand may decide some products never move on price. That is a valid choice, and it is better made deliberately than by default.
- Inconsistency. Two shoppers paying different prices for the same item can feel unfair. Clear, explainable rules help: quantity, stock age and timing are easy to justify; who the shopper is, is not.
- Loss of control. The biggest worry merchants raise is an automated system agreeing to a bad deal. Every rule should be set by the merchant, logged, and easy to switch off.
The common thread is that the merchant decides the limits, and the technology only works inside them.
Five things to do before peak season
You do not need to rebuild your pricing to prepare. Start here:
- Check how you appear in an assistant. Ask ChatGPT for products like yours, with a budget, and see whether you show up and how you compare.
- Clean up your product data. Accurate prices, stock, specifications and condition (for refurbished goods) make you easier to find and easier to recommend.
- Sort your catalogue into fixed and flexible. Most products stay fixed. Pick a small set where a conversation could close a sale you would otherwise lose.
- Set floors from margin. Write down the lowest price you would accept for each flexible product, and what would justify moving it.
- Decide your fallback. Choose what happens when a shopper or an assistant asks for less than your floor: a polite no, or a request passed to you to review.
The price is becoming a conversation
For years, a price was something a store published and a shopper accepted or ignored. As assistants take on more of the searching and comparing, the price is becoming the start of a conversation instead.
The stores that do well will not be the ones with the lowest prices. They will be the ones that have decided, in advance, where they are firm, where they will flex, and by how much.
