See what AI says about you, then get the fixes for every gap.
Recover does two jobs. First, a forensic technical audit that tells you, page by page, what's stopping AI from reading you, and hands you the fix to paste straight in. Second, the AI Mirror Test: it asks the questions your buyers actually ask across ChatGPT, Claude, Perplexity and Gemini and shows you where you're named and where you're invisible. You leave with a clear score and a ranked plan, not a vague to-do list.
The problem
Most businesses have never checked whether AI mentions them. They optimised for Google for years and assumed it carried over. It doesn't. Ranking on Google and getting cited in an AI answer are different games, read off different signals, and the page that wins the search result is often the one the AI ignores.
Only 16% of brands track how they show up in AI search. So the first move isn't to fix anything. It's to find out, precisely, where you stand.
of brands track how they show up in AI search.
Job one
Job one: the technical audit
This isn't a score and a shrug. The audit crawls your site the way an AI answer bot does and reports, page by page, exactly what's holding you back: missing or broken schema, blocked crawlers, robots.txt rules shutting the answer bots out, structured-data gaps, indexability problems.
Then it hands you the fix. For schema, that means a clean block of code written for your page, ready to copy and paste in. No theory, no "consider implementing structured data". The actual markup, on the actual page, with the actual problem named.
Not comfortable pasting code into your site? You don't have to. Done for you ($600/mo) handles it for you.
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Acme Co",
"url": "https://acme.co",
"logo": "https://acme.co/logo.png"
}
Don't want to touch the code? Done for you ($600/mo) handles it for you.
Ask about fixing →Job two
Job two: the AI Mirror Test
Here's where most tools stop short. They check your tags and call it a day. The Mirror Test checks something harder: when a real buyer asks an AI about your category, does it name you?
It works off your profile. Once you've told Recover who you are and who you sell to, it works out the questions your ideal customers are likely to ask an LLM. You can narrow those to specific niches or markets, or, if you're just starting out, ask for a set of foundational questions to get you going.
From that, it generates 50 to 60 questions. You can edit them, add your own, cut the ones that miss. Then it fires every question at each of the four leading models, ChatGPT, Claude, Perplexity and Gemini, and waits for the answers. Fair warning: running that many questions across four models takes a while, so this isn't instant.
When it's done, the results are analysed into a plain picture: which questions name you, which name a competitor instead, and which return someone else entirely. That picture becomes a content gap analysis, the questions you should own but don't, which is the start of your content strategy. Build picks it up from there.
How it works
Point it at your site
The technical audit crawls your live site the way an AI answer bot reads it.
Build your profile
Tell Recover who you are and who you sell to. Narrow to niches or markets, or pick foundational questions if you're new.
The Mirror Test runs
Recover generates 50 to 60 buyer questions, you edit them, then it fires them at all four models. This part takes a while.
Everything gets analysed
Page-by-page technical issues with copy-paste fixes, plus where you're named across the models and where you're not.
You get a ranked plan
What to fix first, in order of impact: technical blockers, then the content gaps the Mirror Test exposed.
What you walk away with
A page-by-page fix list
Every technical issue located on the exact page it lives on, each with the fix written out and ready to paste. Or hand it to the fixing service.
A clear score for where you stand
One honest number for how often AI names you today across the four models. It's the line every future month is measured against. Low to start is normal, and it's a starting line, not a verdict.
Your content gap analysis
The questions you should own but don't, ranked by impact. This is your content strategy on a plate, and it's exactly what Build turns into briefs and drafts.
Then you Build
Recover tells you what's broken and what's missing. Build closes the gap. It refines your voice and your real opinions, then turns that content gap analysis into rich, specific content that gets you noticed next time the AI is asked, the kind that stands out from the generic AI slop everyone else is publishing.
See what Build does →Questions
How long does the audit take?
The technical checks are quick. The full AI Mirror Test takes a while, because it fires 50 to 60 questions across four models and waits for the answers.
Will I need a developer to apply the fixes?
No. The fixes are cut-and-paste, written for your specific pages.
What does the audit actually check?
Your technical readiness, schema, crawler access, robots.txt and indexability, plus, on full access, where AI mentions you across the four models.
What if I don't want to do the fixes myself?
Done for you ($600/mo) handles it all for you.
What's included in the free audit versus full access?
See the full breakdown on the Pricing page.
See where you stand today.
Free technical audit and AI Mirror Test. No credit card, just a clear picture of where you stand.