Generative engine optimization services
Getting named when an AI answers your buyer’s question. We measure where you stand across every engine, fix what is keeping you out, and rerun the same prompts every month so the movement is arguable.
Generative engine optimization is the work of getting a brand named and cited inside AI generated answers, on ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews. It has three parts: being reachable by the crawlers those systems use, being quotable once reached, and being corroborated by sources the model already trusts.
- Engines tested
- 4
- Checks per page
- 36
- Baseline
- Free
- From
- $99/mo
ChatGPT, Claude, Gemini and Perplexity, per prompt
run against your live response, not a staging copy
the audit that scopes the work costs nothing
month to month, no minimum term
Why brands stay invisible to generative engines
In almost every audit we run, the reason is one of these three, and they are not equally common. The first is by far the most frequent and the cheapest to fix.
The crawler never got in
GPTBot, ClaudeBot, PerplexityBot and Google Extended are each allowed or refused separately in robots.txt, and a default WordPress or Shopify install refuses several of them without anyone deciding to. A model cannot cite a page it was never served.
The page arrived empty
Most AI crawlers do not execute JavaScript. On a React, Vue or headless build that renders content client side, what they receive is a shell with a loading state in it. The page looks fine to you and is blank to them, which is why this one takes an audit to find.
Nobody else says it about you
Generative answers lean on third party corroboration far more than classic search does. A brand described only on its own site reads as a claim; a brand described on directories, review sites and in press reads as a fact. This is the slowest of the three to fix and the one that decides the ceiling.
What a GEO engagement actually ships
Named artefacts, in the order they normally land. Nothing here is an activity that shows up on a report without something changing on your site.
A measured visibility baseline
Your real buying questions, asked of each engine, with the answer recorded: are you named, is a competitor named instead, and which sources did the answer draw on. This is the number every later claim is measured against.
Crawler access, opened per bot
robots.txt corrected bot by bot with the reasoning written down, because allowing every crawler is a decision with tradeoffs and should be made rather than defaulted into. Plus llms.txt where it earns its place.
The render gap closed
Server side rendering, static generation or prerendering for the templates that carry your content, so a crawler that runs no JavaScript still receives the words. Usually the single highest impact fix on a modern stack.
Answer shaped content on your live pages
Direct answers in the opening lines, questions as headings, facts with sources attached, published into your CMS by us or handed to your editor. Written so a model can lift a clean passage without inventing the missing half.
Entity and schema work
Organization, Service and FAQ schema, and a sameAs set covering every profile you genuinely own, so an assistant can tell that the brand named in a third party source and the brand on your site are one company.
Corroboration where the models read it
Directory and marketplace listings, review platform presence, and digital PR pitched to publications in your space. Reported as what actually landed, because this is the one line nobody can guarantee.
How the work runs, month by month
Four repeating steps. The first one you can run yourself, right now, without talking to us.
- 01
Measure before touching anything
Full audit across 36 checks plus the prompt set, so both sides start from the same numbers. If the audit says your GEO problem is actually a crawler problem, we would rather say that in week one than sell you content for six months.
- 02
Clear the mechanical blockers
Crawler access, rendering, sitemaps, canonicals and schema. These are the fixes with a measurable before and after inside days rather than months, so they go first and they prove the engagement early.
- 03
Publish what the engines can quote
Answer first pages against the questions your prompt set showed you losing. Every piece is scored on extractability and attribution by the same engine we sell, before it goes out.
- 04
Rerun and report
Same prompts, same checks, same day of the month. Movement is tied to the work that period, and where a movement has no work behind it we say so instead of claiming it.
How you will know whether it worked
The honest answer is that GEO has a measurement problem the whole category is quiet about: generative answers are not deterministic. Ask the same question twice and you can get two different answers, with two different sources. Anyone showing you a single screenshot as proof is showing you one roll of a die.
So the unit here is the prompt set, not the screenshot. The same questions, asked of the same engines, at the same cadence, over months, produces a rate rather than an anecdote, and a rate is the only thing that survives the randomness. That is what goes in the monthly report and what you can rerun yourself between reports.
The second measure is the one nobody can argue with: the audit score on your own pages. Crawler access, rendering, schema and page structure are deterministic, and every one of them either improved or did not. That half of the work is checkable by you, by us and by any agency you ask for a second opinion.
What we will not do is promise a position. There is no ranking to hold in a generative answer and any agency implying otherwise is describing something that does not exist. What we charge, and what the rest of the market charges, is published rather than quoted, for the same reason.
It is fair to ask whether anyone in this category holds themselves to it. We checked: we audited 98 agencies selling GEO with the same engine we run on your site, and a third of them publish pages carrying no date and no named author, which are two of the things they sell the fix for. Our own pages are excluded from that study and scored separately, which the methodology says in as many words.
Find your own blockers first, for nothing
The three checks below find the two most common causes of generative invisibility in about a minute each. Run them before you talk to us or to anybody else.
Or run the same interrogation by hand: what generative engine optimization actually is, getting named in ChatGPT answers, ranking a website inside ChatGPT. Free, no account, and the honest version of the do it yourself option.
What buyers ask about GEO
What is generative engine optimization?
The work of getting a brand named and cited inside AI generated answers rather than only in a list of blue links. It covers crawler access, page structure that a model can quote, entity clarity so the model knows who you are, and third party corroboration so it trusts the claim. The measurement unit is a prompt set rather than a keyword position. Our full explainer covers where the term came from and what the founding research actually found.
How is GEO different from AEO?
GEO is about being named in a generated answer at all, on any engine. AEO is narrower and older: being the specific extracted answer to a specific question, which includes featured snippets and voice results as well as AI ones. Most engagements need both, and we run them as one workstream. Our AEO page covers that half in detail.
How long before anything moves?
The mechanical half moves in days to weeks: once crawler access and rendering are fixed, the next crawl sees a different site and the audit score changes immediately. The citation half compounds over three to six months, because it depends on other people publishing things about you. Anyone promising AI citations in week two is describing a purchase, not an outcome.
Can you guarantee we get cited by ChatGPT?
No, and neither can anybody else. Generative answers are probabilistic, the same question can return different sources on two consecutive asks, and no vendor has a lever that forces an inclusion. What can be guaranteed is that the reasons you are currently excluded get found and fixed, and that the rate is measured the same way every month.
Do we need to change our whole website?
Almost never. The fixes that matter most are template level: how pages render, what robots.txt allows, what schema the templates emit, and how the opening of a page is structured. Those reach every page at once. Rewriting individual pages comes after, and only for the questions the prompt set shows you losing.
What does GEO cost?
Agency retainers in this category run from $2,500 to $50,000 or more a month. Ours are $99, $299 and $799 a month, month to month, because the diagnostic half is automated rather than billed by the hour. The full market table with named agencies and sources is on what GEO costs.
We already have an SEO agency. Does this conflict?
Usually not. The common split is that they keep strategy and content while we take the technical and AI visibility layer, because that is the part most agencies have no engine for. We work to their brief where there is overlap, and everything we change is recorded so nobody is guessing who did what.
Start with the measurement, not the retainer
The free audit gives you the crawler access, rendering and schema findings any GEO proposal would open with, measured on your live pages. Take it to us or to whoever else you were considering.