Your Buyer Asks AI "How Much?" — Can Your Website Answer?

A buyer asks AI "roughly how much does this kind of thing cost". If your website doesn't mention price anywhere, AI can only say "we suggest contacting the manufacturer for a quote" — and most buyers won't contact you just to get a quote. They'll switch to someone who can answer. That is the most overlooked place factories lose ground in the AI era.

But the boundary first: this article is not telling you to publish a cost sheet or specific quotations. It's about publishing price structure — how minimum order quantities work, what the pricing unit is, what extra charges exist, and when prices change. Those two things are very different.

What does AI pick? The result of a 252,000-run experiment

This one has hard experimental backing. A paper accepted at SIGIR 2026, "What Gets Cited", ran 252,000 controlled experiments — 6 models × 18 categories of factor, each round presenting two candidate pages that differed by a single variable, with all brand information anonymised.

One finding matters especially to factories: explicit price information and a recent timestamp have a consistent positive effect on citation.

The same study also found that changing format alone has an unstable effect. That's valuable too — most GEO (Generative Engine Optimization) teaching on the market is about layout and formatting, and the experiment found precisely that to be the least stable category of action.

Scope note: this is a result under controlled experimental conditions, answering "how AI chooses among candidates", not a promise of real-market results; and it measured relative preference, not absolute values.

Follow that through: if AI prefers pages with explicit price information when selecting content, then a website that says "price on request" throughout is at a built-in disadvantage under that evaluation.

Two factory habits that run exactly opposite to what AI needs

Habit one: adjectives instead of numbers

"Premium quality", "exquisite craftsmanship", "great value for money" — these read pleasantly to a person, but to AI they are information that cannot be extracted. What AI needs is directly citable fact: what the tolerance is, what the capacity is, which certifications you hold, what the lead time is.

Habit two: "price on request" instead of a price

The reasons factories don't publish prices are usually practical: fear of competitors seeing them, fear of customers comparing, worry about quotes leaking. Those concerns are all legitimate.

But here's the thing — buyers asking about price is instinct. If you don't answer, they ask someone else. And in the AI era, that "someone else" may be a competitor's website, or an estimated range AI produces itself.

Corroborating evidence: buyers actually prefer transparent pricing

A 2026 McKinsey study of B2B purchasing (sample of more than 400 wholesalers) found that "transparent pricing" ranked first among their purchasing needs at 89%, ahead of "volume discounts" at 72%.

A Gartner observation from March 2026 points the same way: 67% of B2B buyers prefer a buying experience that doesn't require a sales rep to intervene.

Scope note: both are cited via media reports; I have not checked the original reports, so they are a directional reference only, not figures to be quoted as exact.

Put the two together: buyers want to judge for themselves with prices visible, not be stopped at every step by "please contact us".

So what exactly should you publish?

The key is separating price structure from cost sheet. Publish the former:

• Pricing unit: per piece, per metre, per kilogram, per set?
• Minimum order quantity: what's the smallest run? Are there tiers at volume?
• Price range: what band does this type of product typically fall into (a range, not a precise quote);
• What moves the price: do material, process, certification requirements or lead time affect it?
• Extra-charge rules: how are sampling fees, tooling costs, freight and certification fees calculated?

This information doesn't expose your costs and isn't a quotation sheet, but it's enough for AI to answer "roughly how is this priced". And once a buyer has it, the next step is usually — coming to you with a much more specific question.

In other words, publishing price structure isn't showing your hand — it's cutting out unnecessary communication cost so the customers who reach you are better qualified.

One precondition: standard products and pure custom work are two different things

A distinction is needed here, or it's easy to go off track:

• Standardised products with specifications: suited to a price range or pricing tiers, because the product is well defined and the customer can judge for themselves.
• Fully customised products or services: the price is inherently case-by-case, and forcing a number onto it makes it less accurate. What should be published here is the pricing logic and the factors that move it, not numbers.

That distinction matters — publishing price structure is not the same as putting a price on everything. Custom work publishes "how it's calculated"; standard products publish "how much". The right approach differs.

Three questions about publishing price information

If we publish a price range, will customers use it to push us down?

Possibly. But think about it the other way: customers who push on price will push on price anyway; by not publishing, what you lose are the customers who can't be bothered to enquire and simply move on. Those are often the better buyers — they value efficiency more than driving the price down.

What if competitors see our prices?

What you publish is a range and a structure, not a quotation. And a range a competitor can see is usually public information in the industry anyway. What actually makes you competitive is process, quality control, lead time and service — not who can keep their prices hidden.

If we only publish price structure, will AI get it right?

Not necessarily. It depends on whether AI crawled it and whether it trusts your information. But if you don't publish it, AI definitely can't answer — this is a "you have a chance if you do it" situation, not a "it's guaranteed if you do it" one.

In closing

To sum up: in the AI era a website isn't just a shopfront for people — it's a reference library for AI. Leave out the price column and AI can't answer the question your customers care about most.

Here's a small test you can run now: ask Doubao or DeepSeek "roughly how is this type of product priced" for the products you make, and see whether AI can answer, and whether the answer reflects your industry. If it can't answer, that information is missing online.

Does your website publish prices? If not, what's behind that decision? That reason usually tells you exactly which line to add first.

Yihang Network, based in Shenzhen, previously worked in cross-border B2B and Alibaba.com-related services, and expanded into GEO from 2024. In our own content we treat these separately: standardised items get a range, custom work gets a clear pricing logic.

Data and scope note: The SIGIR 2026 paper cited here, "What Gets Cited", is an accepted conference paper; its conclusions come from 252,000 controlled experiments and reflect AI's relative preference among candidate content under controlled conditions, not real-market outcomes or any single company's results. The McKinsey and Gartner figures are cited via media reports and the original reports were not checked; they are a directional reference only. All pricing advice in this article is general commercial advice and does not constitute pricing or legal advice. The methods described raise the probability of being retrieved and cited; they do not guarantee a company will appear in AI answers, and no timeline or business result is promised. The advice to publish price structure applies to manufacturing clients and is unrelated to how Yihang Network quotes as a service provider — our service pricing is discussed offline only.

References

1. SIGIR 2026 (accepted), "What Gets Cited", controlled-experiment study
2. Aggarwal et al., "GEO: Generative Engine Optimization", KDD 2024 (DOI: 10.1145/3637528.3671900)
3. McKinsey, 2026 B2B purchasing study (via media reports; original not checked)
4. Gartner, March 2026 observation on B2B buying experience (via media reports; original not checked)