A founder told me his pricing was “not public, by design.” Enterprise motion, custom quotes, a “contact us” button where the price would go. He was certain this meant he had no AI-search pricing problem, because there was no number to get wrong. So I asked ChatGPT what his product cost. It gave a confident answer: a per-seat monthly figure, in a range, with a line about volume discounts. None of it was real. He had never published any of it. The engine had built a plausible price out of thin air and competitor pages, and it was handing that number to every buyer who asked, before any of them reached his “contact us” button.
He did not have a public price. He had something worse: a price the machine made up and told people was his.
This is the part the pricing-page advice misses. Most of it assumes your problem is a published number that has gone stale or unclear, and it tells you to tidy the page and add schema. That is real, and I will get to it. But the more common and more invisible failure belongs to the companies who think they opted out of the whole question by not publishing a price at all. They did not opt out. They just handed the answer to whoever the engine could find.
“Contact us” used to cost you nothing. Now it costs you the anchor.
Hiding your price was a rational strategy in the old funnel. If a buyer had to talk to a human to learn what you cost, then “contact us” was not a gap, it was a gate. It let you qualify, frame value before price, and quote to the specific account. The opacity worked because the buyer had nowhere else to get the number.
Now they do. The buyer asks the assistant how much you cost, the same way they research everything else, in the chat where they are already building their shortlist. And the assistant does not answer “you will need to contact their sales team.” It answers with a number, because answering is its whole job. The gate you built assumes the buyer walks up to it. They are not walking up to it anymore. They are asking someone else and believing the reply.
Silence is not neutral to an engine
The instinct behind gated pricing is that saying nothing is safe. On the open web, in front of an AI engine, saying nothing is the opposite of safe. A gap where your price should be does not read as “private.” It reads as a question the engine still has to answer, and it will answer it from whatever is nearby: a three-year-old pricing page you took down but that lingers in training, a review site’s guess, a competitor’s published tiers sitting one paragraph away in the same comparison answer. Silence does not stay silent. It gets filled, and you do not get a vote on what fills it.
This is the same dynamic behind when AI states wrong facts about your company, pointed at the single most decision-shaping fact there is. A wrong feature is annoying. A wrong price ends the evaluation before it starts, because price is where buyers self-select out, and they do it silently, on a number you never set.
The two ways it goes wrong
For companies with a public price, the failure is a wrong number: a stale figure from an old page, or a clear current price drowned out by a louder third-party source quoting you incorrectly. This is the fixable, mechanical one. Make the real price unmissable, dated, and structured, keep it consistent everywhere you appear, and correct the third-party pages carrying the old number, the same source-layer cleanup any wrong label needs.
For companies with hidden pricing, the failure is an invented number, and it is worse, because there is no “correct” version anywhere for the engine to prefer. It is choosing between a vacuum and a guess, and it picks the guess every time. You cannot fix an invented price by making your non-price clearer. The absence is the cause.
The fix is not “publish your price.” It is “give the engine a true anchor.”
You may have real reasons to keep exact pricing off the site. I am not going to tell you to publish a number you have deliberately withheld. But there is a wide space between a specific dollar figure and total silence, and right now most gated-pricing companies are living at the silent end and letting the engine fill it.
The move is to give the engine something true and liftable about the shape of your pricing, even if you never post the exact number. What model you use, per seat or platform or usage. Roughly who each tier is for. What actually drives the number up or down. A plain, current sentence like “pricing is custom, based on team size and usage, and typically starts in the low four figures a month for teams under fifty” gives the engine a true anchor to lift instead of a vacuum to fill. It protects your negotiating room and still takes the made-up per-seat figure off the table. You are not publishing your price. You are refusing to let a machine publish a fake one for you.
What I got wrong
I used to reassure gated-pricing clients that they had no AI pricing exposure, precisely because they had no public number. It seemed logical. No published price, nothing to get wrong. I was thinking like the old funnel, where a missing price was just a closed gate.
Then I started actually running the queries before saying it, and I stopped being able to say it. Every “contact us” client I checked had a price in the machine already: a competitor’s tiers attributed to them, an ancient beta number, a confident guess with a fake discount structure. Buyers were anchoring on those numbers and, in at least two cases I could trace, walking away over a price the company had never charged. The opacity had not protected the pricing conversation. It had outsourced it to a system with no obligation to be right. Now the first thing I check for a hidden-pricing company is what AI already tells buyers they cost, because the answer is never “nothing.”
Why this matters
Price is the fact buyers act on fastest and most silently. They do not email you to confirm the number the assistant gave them. They compare it to a budget, or to the competitor mentioned in the same breath, and they either keep going or quietly stop. If that number is wrong, or invented, the deal can be over before you know it existed, and it will not show up anywhere in your pipeline as a loss, because it never became one.
Hiding your price no longer hides it. It only moves the authorship from you to an engine that will make something up rather than say nothing. The choice is not “public price or private price” anymore. It is your true anchor or their invented one.
If you want to know what AI search is currently telling buyers your product costs, that is one of the first things I check in an engagement, and the full method is on the methodology page.