Can AI agents understand your pricing page?

An agent that reads “Plans start at £19” still may not know which plan costs £19, who qualifies for it or whether that is a monthly or annual price. A pricing page can look polished while leaving the details a buyer needs scattered across tooltips, comparison modals and sales footnotes.

The fix is not to write for a machine instead of a person. It is to state the buying terms clearly enough that both can compare plans without guessing.

Put the decision-making facts in the page

For each plan, state the plan name, price, currency and billing interval together. If the price is “per seat,” say what counts as a seat. If a plan is free only for a limited period, name the length of the trial and what happens when it ends. “Contact sales” is a valid choice, but explain who the plan is for and what a buyer can do next.

Then make the limits concrete. “Unlimited projects” may have a fair-use limit or exclude archived projects. “Priority support” should say which channels and response target it includes. If usage overages are possible, show how they are calculated and whether the account needs to opt in.

Important details to keep close to the plan include:

• monthly and annual rates, including the saving or discount calculation;
• currency, billing interval, taxes and any setup fee;
• feature and usage limits, such as seats, storage, projects or API calls;
• trial length, cancellation rules and renewal behavior;
• eligibility restrictions by company size, location or use case;
• add-ons, overages, implementation fees and support terms.

If a fact changes based on location, make the rule visible and test the page with that location selected. Do not expect an agent to infer a price from a screenshot or a calculator whose result never appears in the page content.

Test the questions a buyer asks

Give a test agent specific questions and compare its answer with the pricing page and checkout. For example: “What is the lowest plan for five users if billed monthly?” “Does the trial require a card?” “Can I cancel an annual plan early?” “How much does one extra seat cost?”

For every answer, record the exact page section and the matching value in your billing system. If the agent returns the wrong interval, drops a limit or invents a discount, find where the page leaves room for that interpretation. Run the same questions after a pricing change; an old cache or search snippet should not be mistaken for current terms.

Check the initial HTML response as well as the normal browser view if the price is populated by a client-side widget. Google's Search Central guidance explains its own JavaScript rendering, but agents and crawlers do not all render pages the same way. Keep the price and core terms in accessible page content, then use structured data only when the markup matches those visible details. Google's guide to generative AI Search says new special files or AI-specific markup are not required for appearance in its AI features. See the AI features guide and product structured data documentation.

Treat pricing as a data contract

Your page, billing system, product feed and sales materials should agree. If the page says £19 per month but checkout charges £228 at the start of an annual term, explain that the annual total is billed upfront. Keep a source of truth for the price and limits, and define how and when those values are refreshed.

That does not mean every price belongs in public structured data. Private negotiated rates, account-specific discounts and location-dependent taxes need the right context. Show what is public, label what depends on the account or region, and require a human confirmation before any purchase is made through an agent-supported flow.

What a readiness scan can and cannot establish

A scan can check whether a public pricing page is reachable and whether its key content appears in the response it inspected. It cannot verify what a signed-in customer sees, confirm that billing uses the same limits, or predict whether an AI platform will recommend your plan.

Dedica's free website readiness scan can help find access and content issues on the public page. Pair it with a buyer-question test and a checkout comparison before treating pricing as ready for an agent-assisted purchase.

For measurement, separate those page checks from actual mentions and traffic. The guide to AI agent readiness vs AI visibility explains what each result can establish.

Sources

• Optimizing your website for generative AI features on Google Search, Google Search Central.
• Product structured data, Google Search Central.
• Fix Search-related JavaScript problems, Google Search Central.