Maveriko
Sign in
New audit
Product
Site crawlCompetitor gap reportAI Readiness benchmarkHow scoring works
Free tools
Keyword ideasAI readiness checkSchema codeRobots & sitemapllms.txt fileTitles & meta tagsAI quotabilityAll free tools
Learn
GEO explainedAI visibilityField guideTool comparisons
Pricing

GEO: Generative Engine Optimization

GEO is the practice of structuring content so AI answer engines can read, trust and cite it. Where SEO earns you a ranking on a results page, GEO earns you a quotation inside the answer itself — and the two share a foundation but reward different things.

GEO is the on-page half of AI visibility. This hub covers what it is, how it compares to SEO, and how to audit a page for it — plus playbooks for SaaS, ecommerce and agencies.

Run a free GEO audit →Free · scores your live URL for SEO, GEO & entity authority

SEO and GEO, side by side

SEO and GEO are one technical foundation pointed at two different wins. SEO earns a position in a list of links; GEO earns a sentence inside the answer itself:

Traditional SEOGEO
GoalRank in the resultsBe quoted in the answer
Unit of successA ranking URLA citable passage
Trust signalBacklinksEntity authority, corroboration
Format that winsKeyword-relevant copySchema, direct answers, tables

The two halves of a GEO score

A GEO score is two measurements, not one: what your page gives an engine, and whether that engine already knows your brand. A flawless page from a brand no model recognises is still not recommended:

  • Page readiness — can an engine fetch this URL and lift a citable claim from it? Schema, structure, direct answers, crawler access.
  • Entity authority — does AI already know the brand exists, from Wikipedia, Wikidata and consistent public records? For recommendation questions, engines name brands, not URLs.

In Maveriko’s score the split is 60/40 in favour of the page, with the full weighting on the methodology page.

How an answer engine decides what to cite

Citation is the last step of a five-stage pipeline, and a page can be eliminated at any one of them. Understanding where you drop out is most of the work, because the fixes are completely different at each stage.

  1. Fetch. A crawler — GPTBot, PerplexityBot, ClaudeBot, Google-Extended — requests the URL. If robots.txt disallows it, nothing that follows can happen. This is the single most common total failure, and it is usually accidental.
  2. Render. The fetched HTML is parsed. Most AI crawlers do not execute JavaScript, so a page that ships an empty shell and paints its content client-side is, to them, a blank page. Your visitors see an article; the engine sees nothing.
  3. Chunk. The text is split into passages, usually around headings. A section that answers its own heading in its first sentence survives chunking intact. A section that meanders for three paragraphs before reaching the point gets split away from its answer.
  4. Retrieve. For a given question, candidate passages are ranked by semantic similarity — and, increasingly, by whether the source is a recognised entity rather than an unknown domain.
  5. Attribute. The model composes an answer and names its sources. Passages that read as self-contained factual claims get named; passages that only make sense in context get absorbed without credit.

Stages 1 and 2 are binary — you pass or you are invisible. Stages 3 to 5 are gradual, and they are where formatting work actually compounds.

What GEO is not

GEO is a readiness discipline, not a measurement of outcomes. It is worth being blunt about this, because a great deal of marketing in this category is not.

Three different things get sold under one name. SEO health is whether search engines can crawl and understand a page. AI readiness — GEO — is whether that page is structured to be retrieved and quoted. AI visibility is whether engines actually cite you today, which changes per model, per prompt and per week, and cannot be established by scanning your HTML.

Google’s AI Overviews are the highest-volume case of this today; the AI Overview optimization guide covers how sources are chosen there specifically. Any tool that reads your page and reports a “ChatGPT visibility score” is inferring the third from the second. GEO work raises the probability of citation; it does not measure it. We hold that line deliberately and explain the reasoning on the methodology page.

How to audit a page for GEO

A GEO audit works through the pipeline in order, because a fix at stage 4 is wasted if you fail at stage 1. Run these against any URL you care about:

  1. Check crawler access. Read robots.txt for explicit Disallow rules against GPTBot, PerplexityBot, ClaudeBot, CCBot and Google-Extended. Many CDNs and security plugins block them by default.
  2. View source, not the rendered page. If the body text is not in the raw HTML, any crawler that does not run JavaScript will miss it. This one check invalidates more GEO work than any other.
  3. Validate the JSON-LD. Organization on the homepage, Article on posts, FAQPage where you answer questions, Product where you sell. Broken JSON is worth zero, and it fails silently.
  4. Read your first sentences. Every heading should be answered directly beneath it, in one sentence, in the form “X is Y”. If your opening line is throat-clearing, the passage is unquotable.
  5. Convert comparisons to tables. If you describe a comparison in prose, engines rarely reconstruct it. Real <table> markup is lifted almost verbatim.
  6. Check the entity, not just the page. Search your brand name on Wikidata. If nothing resolves, the off-page half of your score has nowhere to come from, and recommendation-style questions will not name you.

What our benchmark found

These are findings from Maveriko’s own benchmark: 67 named homepages, all measured under one engine on 12 Sept 2026.

  • The median homepage scored 79 for SEO Health but 47 for AI Readiness: a 32-point gap between readiness for search and readiness for AI answers.
  • 62 of 67 (93%) did not open every section with a direct answer.
  • 15 (22%) blocked, in robots.txt, a crawler that AI answer engines depend on, and 13 more restricted them in part, shutting out only the training crawlers or closing some paths.
  • 18 (27%) had no JSON-LD structured data that parsed.

Every row, including each site we could not score and why, is on the AI Readiness Benchmark.

The mistakes that cost the most points

Across the pages we score, the same handful of faults account for most of the lost ground — and none of them are subtle once you know to look:

  • Client-side rendering. A React or Vue app with no server rendering is invisible to non-executing crawlers. Nothing else you do matters until this is fixed.
  • Blocking AI crawlers by accident. A blanket bot rule inherited from a WAF or a plugin, never reviewed since.
  • Thin pages. Under roughly 300 words there is not enough material to chunk, retrieve or quote — a well-marked-up page with nothing on it is still nothing.
  • Schema that describes the wrong thing. WebPage on everything, or a single Organization block copied to every URL, tells an engine nothing it can use.
  • No stated author or date. Engines weight corroborable, attributed material. An undated, unsigned page reads as unowned.
  • An invisible brand. No Wikidata item, no consistent sameAs links, a different company name in every profile.

Common questions about GEO

What is GEO?
GEO, or Generative Engine Optimization, is the practice of structuring content so AI answer engines can read, trust and cite it. Where SEO earns a ranking, GEO earns a quotation inside the answer itself.
Is GEO the same as AI visibility?
GEO is the on-page half of AI visibility. Visibility is the outcome — being cited; GEO is one of the inputs, the work of making a specific page structurally quotable. Entity authority is the other input.
How do I audit a page for GEO?
A GEO audit checks retrievability, structured data, extractable formatting and entity authority, then scores how ready the page is to be cited. You can run one free on any URL from the Maveriko homepage.
Does GEO replace SEO?
No — GEO sits on top of SEO and depends on it. An answer engine still has to crawl, parse and index the page before it can quote anything, so the technical foundation is shared. What changes is the target: SEO optimises for a position in a list, GEO optimises for a passage an engine is willing to lift and attribute. Sites that abandon SEO fundamentals to chase GEO end up failing both.
How often should I re-audit a page for GEO?
Re-audit after any change to your rendering, robots.txt, CDN or schema, because those are the changes that silently break retrievability. Beyond that, a quarterly pass is enough for stable content. The signals GEO measures are structural, so they do not drift on their own the way rankings do — they change when you change something.
Can I do GEO without changing my website?
Only partly. Entity authority — the off-page half — is built off-site through Wikidata, consistent profiles and public records, so it can improve while the site stands still. Page readiness cannot: schema, direct-answer openings, tables and crawler access all live in your HTML. Most sites find the off-page half is the slower of the two but the one they have neglected entirely.

Sources

Published · Last reviewed .

Read the cluster

7 guides, from the definition to industry-specific playbooks.

Run a GEO audit on any URL

Paste a link and get its SEO, GEO and entity-authority scores in about 15 seconds, with the exact fix for every gap.

Run a free GEO audit →