SecondDoor

Pricing an AI agent can actually read

Why prices that need JavaScript are invisible to buyers that do not run it, and what to publish instead.

The test

Fetch your own pricing page with curl and read what comes back. If there are no prices in it, an agent evaluating you cannot compute what you cost.

This is more common than it sounds. Pricing tables are frequently rendered client side from a config service, which means the numbers exist in an API response rather than in the HTML. A person sees prices. A fetch sees an empty shell.

What to publish

Put the numbers in the server-rendered HTML. If your pricing is genuinely dynamic, render a sensible default server side and let the client refine it.

Then add Product and Offer structured data with the prices and currency. That gives a machine a parseable source rather than a number it has to scrape out of a table cell, and it is the difference between the middle and the top band of this dimension.

Quote-gating is a decision, not an accident

Contact-sales pricing is a legitimate strategy, and this is not an argument against it. It is an argument for knowing what it costs you: an agent asked to compare three vendors and pick one on price cannot include you in the comparison.

If enterprise is quote-gated, publish everything below it, and publish the limits. Total cost that cannot be computed is the same as no price at all.

Questions

Is a pricing page with a calculator good enough?
Only if the underlying numbers are in the HTML. A calculator that fetches rates after load is invisible to anything that does not run scripts.
What about prices that vary by region?
Render one region server side and mark it up as structured data. A machine that can read one currency confidently is better served than one that can read none.
How pricing legibility is scoredCheck your own site
Pricing an AI agent can actually read, SecondDoor