Your Pricing Page Says "Contact Us." AI Answer Engines Read That as "No Answer."

A prospective client opens an AI assistant and types the question they were embarrassed to ask you on the phone: how much does a dental implant cost in this city, what does a wedding photographer charge for eight hours, what is a reasonable monthly fee for bookkeeping at a business with three employees. The assistant answers with a range. It names two or three businesses that published one. Yours is not among them, because your pricing page says "Every project is different. Contact us for a quote." You did not lose that prospect to a competitor with better work. You lost them to a competitor who typed a number into a web page.

What an answer engine actually does with a pricing page

AI answer engines are not search engines with a chat box on top. When they answer a cost question, they look for pages that contain a cost. They extract the figure, the unit it is attached to, and any conditions around it, and they compose an answer from the pages that gave them something to extract. A page with no figure contributes nothing to that answer, so it is not cited, so the business behind it does not appear.

This is different from the classic search result. In a Google results page, "Contact us for a quote" could still rank, because the page ranked on its heading and its links, not on whether it answered the question. The answer engine has moved the bar. It is grading the page on whether it contains the answer, and "contact us" is a polite way of saying it does not.

Why the number gets hidden in the first place

The reasons business owners give for hiding pricing are real, and it is worth taking them seriously before arguing against them. The first is that every job is different, so a single number would be wrong for most clients. The second is that competitors will read it and undercut it. The third is that a number on the page ends the conversation before the salesperson has a chance to explain the value. The fourth is that a low number attracts the wrong clients, and a high one scares off the right ones.

Every one of these is an argument against publishing a single price. None of them is an argument against publishing a range with conditions. The prospect who reads "most kitchen remodels we do fall between the low and high figure, and here are the three things that move a job toward either end" has been given an answer, has been given a reason to talk to you, and has been given a way to disqualify themselves if they were never going to afford you. The competitor who reads it learns nothing they could not have learned by phoning you with a fake name.

The four things an answer engine can extract

A pricing page that works for answer engines contains four things, and it contains them in plain text on the page, not inside an image, a PDF, or a form that reveals the price after an email address is entered.

• A range, with both ends stated. "From" prices are half an answer and engines treat them as such.
• The unit. Per hour, per session, per month, per square metre, per project. A number without a unit gets quoted with the wrong unit, or not at all.
• The conditions that move the number. Two or three, in words. Size, urgency, location, materials, how many people are involved.
• What is included and what is not. The single most common follow-up question in the AI transcript is "does that include X", and if the page already answers it, the page gets cited for that too.

A structure you can copy

The pattern below is the one that has held up across the service businesses we have applied it to. It is not a design; it is an order of information, and it works on a plain page with no styling at all.

1. One sentence stating the range and the unit, before any explanation. This is the sentence that gets extracted.
2. A short paragraph on what most clients actually pay, and where in the range they land. This anchors the number without committing to it.
3. A list of the factors that move the price, each with one line on which direction it moves it.
4. A list of what is included at each end of the range, and what is always extra.
5. The call to action, which is now "get your exact number" rather than "find out if we are affordable".

Notice what is not on the list. There is no tier table with three columns and a highlighted middle option unless you genuinely sell three fixed packages. Invented tiers are worse than a range, because the engine extracts the tier prices as if they were real products, and the first conversation with the client starts with you explaining that the package on the page does not exist.

The objection about competitors, handled honestly

The competitor who undercuts your published range has told the market they are cheaper than you. If the work is genuinely comparable, you were going to lose that client anyway, on the phone, after spending an hour on a site visit. If the work is not comparable, the published range is now the thing that separates you: yours has conditions and inclusions attached, theirs has a lower number and nothing else. The client who reads both learns that one of you thought about their job and one of you thought about winning the click.

The pricing page is not where the sale is closed. It is where the wrong prospects leave and the right ones decide to call. Hiding the number does not stop that sorting from happening. It just makes it happen somewhere you cannot see.

What to check after you publish

Two weeks after the page goes live, ask the question yourself, in the three or four answer engines your clients actually use, in the words a client would use. If the page is being cited, you will see the business named and the range quoted. If it is not, the usual reasons are that the number is inside an image or a table the engine did not parse, the page is blocked from crawling, or the page took long enough to load that the engine gave up. Each of those is a fix, not a strategy problem.

If you want to know whether your pricing page, and the rest of your site, is being read the way an answer engine needs to read it, the free audit at netwebmedia.com checks exactly that, along with the schema and the crawl signals that decide whether a page is cited at all.

← Back to all articles