Can you guarantee ChatGPT or Perplexity mentions?

No, and anyone who does is selling you something. AI answer engines change which sources they cite week to week, and citations depend on prompt phrasing, retrieval freshness, and trust signals outside your control. What we can do is significantly improve the signals that make citation more likely — and measure the change.

Is AI SEO different from traditional SEO?

It overlaps on hygiene (crawlability, schema, freshness, link reputation) but diverges on intent. Traditional SEO optimizes for ranking in a list of links. AI SEO optimizes for being the source quoted inside a single-paragraph answer. The work that pays off in both: clear entities, structured data, unique first-party data, and short quotable lines.

What is generative engine optimization (GEO)?

GEO is the term for optimizing content to be cited by generative AI answer engines like ChatGPT, Perplexity, and Google AI Overviews. It is a sub-discipline of SEO that focuses on entity clarity, citation-friendly formatting (definitions, lists, comparison tables), and trust signals (external mentions, schema, freshness). Most of what works for SEO still works for GEO; the differences are mostly about how content is structured for excerpting.

How do you measure AI search visibility?

We test a set of prompts representative of your buyer's journey — category queries, comparison queries, branded queries — across ChatGPT, Claude, and Perplexity. For each prompt we record whether your brand appears, where it appears (cited, mentioned, omitted), and what competitor is in the position you want. Baseline plus 30-day re-test gives you a real signal.

What is llms.txt and do we need one?

llms.txt is a one-file convention that gives AI crawlers a clean, citation-ready summary of what your site is, which pages matter, and how to cite you. We recommend every site have one — implementation is ~20 minutes, and it improves the quality of citations even when it does not change the quantity. See our guide on llms.txt for the full spec.

AI SEO Audit for Google AI Overviews, ChatGPT and Perplexity

More than 30% of category-defining queries now start in ChatGPT, Claude, or Perplexity — and end without a click. Our AI SEO audit shows what those answer engines currently say about your brand, where you are missing from answers your competitors appear in, and the highest-leverage changes that close the gap.

Why an AI-specific audit, now

In Q1 2026 about a third of category-defining queries that used to land on a list of ten blue links now end without a click — the user got an AI-generated answer and stopped searching. That answer cites somebody. If it is not citing you, you are losing a share of consideration you used to win for free.

Classic SEO audits do not measure this. They measure rankings, not citations. They measure CTR from result pages users no longer see. An AI SEO audit measures the new surface: which prompts you are absent from, which competitor is in your spot, and what is in their content that is not in yours.

What we analyze

Prompt visibility

We test a set of prompts that span your buyer's journey — category queries ("best X for Y"), comparison queries ("X vs Y", "alternatives to Z"), branded queries ("is X any good", "X reviews"), and use-case queries ("X for [specific scenario]") — across ChatGPT, Claude, and Perplexity. We record: did your brand appear, where (cited, mentioned, or omitted), and which competitor is in the spot you wanted.

Entity clarity

AI engines build a model of who you are from many signals. If those signals are noisy — your brand name overlaps with a common word, your category description differs across pages, your founding facts vary between LinkedIn and your About page — the engine hedges or picks a different cleaner-entity competitor. We audit your entity surface: schema, Wikidata, llms.txt, the first paragraph of your homepage, your About page.

Source-worthy content

AI engines prefer to cite specific over generic. Pages that contain unique numbers, original research, named comparisons, and concrete examples get cited; pages of definitional content that any competitor could have written get summarized without attribution. We map your content library against this filter and identify which pages are citation-ready and which are silently being summarized.

External signals

Wikipedia / Wikidata presence and accuracy
Mentions in trusted secondary sources (TechCrunch, G2, Reddit, category-leading newsletters)
Quote-friendly language patterns in your own copy — short, declarative sentences with concrete numbers excerpt better than long hedged ones
Consistency of brand name and category description across owned and earned media

Why AI search is different

Two structural differences matter:

Full-question intent.Classic SEO indexes keyword-style queries ("crm software comparison"). AI search receives full-sentence questions ("which crm is best for a 12-person agency that also needs project billing"). The pages that get cited are the ones that answer those long questions directly, often in the first paragraph.
Re-ranking, not ranking. AI engines fetch a small set of candidate sources via traditional search, then re-rank them for excerptability. Being in the first ten results is necessary but not sufficient; the re-rank step is where most of the leverage now sits.

What improves AI visibility (in priority order)

Entity disambiguation. If your brand name overlaps with anything else, fix that first. Schema, Wikidata, llms.txt, and the first paragraph of your homepage all need to agree on what you are.
Specific-over-generic content. Add concrete numbers, named comparisons, and original data to your pillar pages. Generic definitional content gets summarized away.
FAQ structure and schema. Questions and answers, short and structured, are the format AI engines lift verbatim. FAQ schema doubles the signal.
Topical depth. Single article on a topic loses to ten interlinked articles. Build out the hub if you want to be treated as a category authority.
External mentions. Earned mentions in trusted sources move the needle on which entities get cited. This is slow work but compounds.

What you get back

Prompt map. The 15-30 prompts we tested, your appearance status on each, and the competitor in the spot when you are absent.
Visibility baseline. Your current cited-mention share across ChatGPT, Claude, and Perplexity for those prompts.
Entity audit. A clear list of contradictions across your owned signals (schema, llms.txt, homepage, About, Wikidata) and how to resolve each.
Content opportunities. The 3-5 content gaps most likely to close the citation delta, with format recommendations and the prompts they would address.
30-day re-test.We re-run the same prompts in 30 days and document the lift. No subjective "trust us" reporting.

Start with the free structural audit — it includes a baseline AI-search read on your domain. Submit your URL and you will see the AI-search section in the first email. The deep audit goes further and the bespoke engagement goes deepest.

An AI SEO audit for the search journeys buyers are using now.