How do you optimize a site for AI search?

Not with keywords. An answer engine has to be able to fetch the page, resolve the entity behind it and find the claim in server-rendered HTML. Most sites fail on one of those three before ranking is even a question, which is why the work starts with the crawler policy, the schema graph and the identity files.

Is this just SEO under a new name?

The overlap is real and the failure modes are different. A page can rank in Google and still be invisible to an engine that cannot resolve its schema or never reaches the content because it arrives after JavaScript. Those are the failures this fixes.

$1,500, fixed. Before that the analyzer at /analyze measures the site for nothing, so you can see what is actually broken before deciding whether it is worth repairing.

How do we know it worked?

The same checks run again after the repair, and the analyzer keeps re-measuring on its own. A visibility engagement that cannot be re-run is an opinion, not a repair.

All service tracks

AI search optimization

Your buyers ask an assistant before they ask a search box. If the assistant cannot retrieve your pages, cannot work out which company you are, or has nothing worth quoting, you are absent from the answer it gives. We measure that first and fix what the measurement found.

Fee: $1,500
Length: 1–2wks

What the engagement covers

Scope of work4

01Open the door to the crawlers that answerGPTBot, ClaudeBot, PerplexityBot and the rest each read robots.txt on their own terms. We audit what your policy actually allows against the crawlers that matter, rather than against a template somebody copied.
02Make the identity resolvableAn assistant that cannot tell which company you are will not name you. llms.txt, the Schema.org graph and the identity files get written so the entity behind the site reads the same way twice.
03Fix the pages an engine cannot retrieveContent that only exists after the browser runs is content an engine has nothing to quote from. We find the pages that never reach the server HTML and say which are worth moving.
04Measure the same thing twiceThe audit that opened the engagement runs again at the end, on the same checks. Anything we claim to have changed is a number visible on both reports.

What you get out of it

A crawler policy written against the assistants that answer, not a template
llms.txt, the schema graph and the identity files in one consistent state
The list of pages an engine cannot retrieve, with the reason for each
Opening and closing audits side by side, run on the same checks

Where this sits7

Visibility AnalyzerfreeFree
Free AI DiagnosticfreeFree
Managed Automation$750/mo
Visibility Upkeep$900/mo
Visibility Sprintyou are here$1,500
Discovery Sprint$2,500
Implementation Sprintfrom $5,000

Two ways to close what the audit found

The report is yours whichever you pick, and every finding in it names the file or the page it came from. What differs is who does the work, and whether anyone measures again afterwards.

Run it yourself

Most findings are an afternoon for whoever owns the site, because the report says which file to open
The crawler policy is the one to check twice: what robots.txt allows and what the CDN does to the request are different questions
The schema graph and the identity files have to agree with each other, or the entity reads as two companies
Pages that only exist after the browser runs are a rendering decision, not a copy one
Then measure again on the same prompt set, because a repair nobody re-measured is an opinion

Visibility Sprint

Have it done

The same repairs, inside one to two weeks rather than between other work
Opening and closing audits on the same checks, so anything we claim to have changed is a number visible on both
Nothing in the plan is withheld if you read it and decide to do it yourself

Not for you if

Where this is the wrong buy3

You need it to have worked by the end of the month. The technical repairs land in days and the closing audit shows them; whether a model starts naming you is slower than that, and nobody here will put a date on it.
You want volume. This publishes nothing on your behalf, and no amount of publishing repairs a crawler policy that excludes the crawler.
There is nothing on the site worth quoting yet. An engine that can reach a page still needs a claim to lift from it.

Asked before anyone buys this

How do you optimize a site for AI search?: Not with keywords. An answer engine has to be able to fetch the page, resolve the entity behind it and find the claim in server-rendered HTML. Most sites fail on one of those three before ranking is even a question, which is why the work starts with the crawler policy, the schema graph and the identity files.
Is this just SEO under a new name?: The overlap is real and the failure modes are different. A page can rank in Google and still be invisible to an engine that cannot resolve its schema or never reaches the content because it arrives after JavaScript. Those are the failures this fixes.
What does it cost?: $1,500, fixed. Before that the analyzer at /analyze measures the site for nothing, so you can see what is actually broken before deciding whether it is worth repairing.
How do we know it worked?: The same checks run again after the repair, and the analyzer keeps re-measuring on its own. A visibility engagement that cannot be re-run is an opinion, not a repair.

Next step

Put your own visibility on the table

One to two weeks: crawler policy, the schema graph and the identity files put right, then the same audit run again so the change is measured. Thirty minutes first, to work out whether that is what your site needs.

Book a call about visibility

Thirty minutes, free.

Fee: $1,500
Length: 1-2 weeks

Re-audited at the end against the same checks, so the change is measured rather than claimed.

Or run the free visibility audit first