Get recommended by AI

Someone asks ChatGPT for a company like yours, and the answer names three or four. This is the work that puts yours in a position to be one of them, and the measurement that says whether it is.

Measured every month, in Spanish and in English

People ask an AI assistant for a recommendation and our name never comes up.

Almost none of these start as a digital project. They start as a sentence somebody says in a meeting, and they end in an answer that already named a few companies before this one was in the conversation.

An assistant answers with what it can read and attribute. When a site hands it structure, content and technical signals, the machine repeats what the business does without inventing the rest.

  • When it does name us, it says something that isn't what we do.
  • The site looks good, but nobody does anything on it.
  • The people who find us are the ones who already knew our name.

Three jobs, one on top of the other.

Measuring without touching the site leaves a number that never moves. Touching the site without measuring leaves work nobody can check. The two run together, on the same list and the same date every month.

01

The questions people actually type

A fixed list of the queries your market actually types, in Spanish and in English: the ones that describe a problem the way the person who has it would describe it, and the ones that ask for a category. It gets written before the answers come back, it does not get edited afterwards, and it runs the same way every month with the date written next to the result.
02

A site a machine can read

Pages a crawler reads without running JavaScript, structured data that says who the company is and what it does, and a robots.txt that lets the AI crawlers in instead of leaving them out by default. This is the part nobody sees and the part that decides whether there is anything to quote.
03

The pages that answer

One page per problem, with the problem in the title and the answer in the first paragraph, in the language the question was asked in. That is the shape a machine can quote without rewriting it into something else.

The number is handed over the way it came out, the month it climbs and the month it sits still. What gets promised is the same list, the same method and the date in plain sight.

What you take with you.

Scope gets decided on the call: the list for a company selling in one country is shorter than the list for one selling in three, and the technical work depends on how the site that already exists was built.

  • Your list of queries, in both languages, with the day-one number written down.
  • The technical work underneath, done on your site: pages a crawler reads without running JavaScript, structured data, and a robots.txt that lets the AI crawlers in.
  • Pages that answer the problem in their first paragraph, in the language the question was asked in.
  • The next measurement, with the same list, the same method and the date in plain sight.
  • The number, however it comes out.

Pura Vida Growth did this same work on its own site before selling it to anyone. Problems we solve is what that looks like from the outside.

Want to see your own case first? Does AI read you? is free and takes a minute. Some people call this AEO, GEO or answer engine optimization. We prefer to say what it does.

What usually comes next

Let's start with your list.

Thirty minutes to look at your case: which queries describe the problem your customers have, in which languages, and what your site hands a crawler today.

Book the call