What is GEO, and how is it different from SEO?

Generative Engine Optimization is the practice of getting a brand named and cited inside AI-generated answers such as ChatGPT, Perplexity, Gemini and Google AI Overviews, rather than ranked in a list of links. It overlaps with SEO on crawlability, depth and authority, but optimises passages to be quoted rather than pages to be clicked, and weights third-party corroboration far more heavily.

What is llms.txt and does it matter?

It is a plain-text file at a domain root that summarises what a business does and points to canonical answer pages in a format aimed at language models. Adoption across AI providers is still uneven, so it is a cheap and worthwhile addition rather than a strategy in itself.

Why do reviews and third-party roundups matter so much for AI visibility?

When an assistant is asked to recommend a company, it synthesises independent sources that have already made that judgement, such as review platforms, trade roundups and comparison articles, rather than weighing marketing copy. A brand's own site establishes what it does and charges; third parties establish whether it is any good. Together, proof and authority account for 30 of the 100 points in this AI visibility checker.

How often should a brand reassess its AI visibility?

Quarterly. Technical and content signals move within weeks of a focused sprint, while entity strength, review proof and earned media move over quarters. Citation share across a fixed prompt set is worth tracking more frequently than the AI visibility score itself.

What is the fastest way to improve a low GEO score?

Confirm that AI crawlers such as GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot are not blocked in robots.txt, a CDN or a firewall, and view your page source with JavaScript disabled to check the content is actually present. After that, publish one honest comparison page covering your brand and its three closest competitors, including the scenarios where a competitor is the better choice.

How is the AI visibility score calculated?

Fifteen questions are grouped into six weighted pillars: answer-ready content 24 points, structured data and machine access 18, entity strength 16, third-party proof 16, authority and earned media 14, and technical measurement 12. Each answer awards 0, 35, 70 or 100 percent of that question's weight, and the total is expressed out of 100 alongside a fix list ordered by points recoverable.

15-Signal Audit · 3 Minutes · No Login

AI Visibility Checker: Score Your Brand Across ChatGPT, Perplexity, Gemini & AI Overviews

Q: Is an AI visibility checker the same as an AI visibility audit?

No. An AI visibility checker measures whether the signals that drive citation are present on your brand. A full AI visibility audit queries the assistants directly across a defined portfolio of buyer prompts and measures how often you are actually named, in what context, and which competitor is named instead. The checker explains why you are losing; the audit quantifies how much you are losing.

Q: Does GEO work for local and multi-location businesses?

Yes. Assistants answering location-based prompts rely heavily on Google Business Profile data, review volume and recency, and consistent business information across directories, which local businesses can usually improve faster than national brands can.

When a buyer asks ChatGPT, Perplexity, Gemini or Google’s AI Overviews to recommend a company like yours, a specific set of signals decides whether your name comes up or a competitor’s does. This AI visibility tool grades your brand against all fifteen of them, weights them by how much they actually influence citation, and returns a 0–100 score with a fix list ordered by the points each change is worth.

15 questions across 6 weighted pillars
Prioritised fix list, ranked by points recoverable
Nothing to sign up for — your score updates as you answer

Start the check See our GEO services

Answer-Ready Content

24 points

Do you publish comparison, alternatives and “best of” content about your own category?

Pages like “X vs Y”, “alternatives to [competitor]”, “best [category] companies”. These are the single most-cited content type in AI buying answers.

Is your content structured so a single passage can be lifted and quoted?

Question-shaped headings, a direct answer in the first two or three sentences beneath each one, definitions, tables and lists — rather than long narrative build-up.

Is your pricing information available on your website?

Assistants routinely decline to recommend brands they can’t describe commercially, and pricing is the most common question in buying prompts.

Structured Data & Machine Access

18 points

How complete is your schema markup?

Structured data is how a machine reads your business as facts rather than as prose it has to interpret.

Do you have an llms.txt file?

A plain-text map at your root that tells language models what your business does and which pages carry the canonical answers.

Have you verified that AI crawlers can actually reach your site?

GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended. Many sites block these by accident through a plugin default or a firewall rule.

Entity Strength & Consistency

16 points

Is your business described identically everywhere it appears online?

Name, category, locations, services and the one-line description across your site, LinkedIn, directories, review platforms and social profiles.

Does your brand exist in the structured knowledge sources models train on?

Wikidata, Crunchbase, LinkedIn company page, industry association listings, established sector directories.

Is your content attributed to named, verifiable experts?

Real bylines with credentials, biography pages and links out to the author’s professional profiles.

Third-Party Proof

16 points

What does your review footprint look like?

Google, G2, Trustpilot, Clutch, Capterra or whichever platforms your category is judged on. Volume and recency both count.

Does your brand appear in third-party “best [category]” roundups?

The listicles published by trade media, review sites and independent bloggers — the pages AI assistants lean on most heavily when asked to recommend.

Authority & Earned Media

14 points

Have you earned mentions in publications that AI models treat as credible?

Trade press, respected industry publications, mainstream business media, well-known independent analysts.

Do you publish original data, research or benchmarks?

Proprietary statistics are disproportionately cited, because they exist nowhere else and models have no substitute source to reach for.

Technical & Measurement

12 points

Does your content appear in the raw HTML, without JavaScript?

Most AI crawlers do not execute JavaScript. If your copy is rendered client-side, they see an empty page.

Are you measuring your visibility inside AI assistants?

You cannot improve a channel you have no reading on, and AI referral traffic is invisible in default reporting.

0out of 100

Not started

0 of 15 answered

Answer the fifteen questions on the left. Your score, your weakest pillars and a prioritised fix list appear here as you go — and your first three answers usually tell you most of the story.

Pillar breakdown

Answer-ready content0 / 24

Structured data & access0 / 18

Entity strength0 / 16

Third-party proof0 / 16

Authority & earned media0 / 14

Technical & measurement0 / 12

Your priority fixes

Your fix list builds itself as you answer — ordered by how many points each change puts back on the board, not by how easy it is.

Get the fixes done for you

Free 30-minute call · We’ll show you which prompts you’re missing from

This is a self-assessment, not a live audit. It measures whether the signals AI systems use to select and cite sources are present on your brand — it does not query ChatGPT, Perplexity or Gemini directly. Two brands with identical scores can still have different citation share, because competition inside your specific category matters as much as your own visibility. Treat the score as a diagnosis of your foundations and the fix list as the order to work in.

Methodology

Why these fifteen, and why these weights

GEO checklists usually give every item equal billing. That’s how brands end up with perfect schema, an llms.txt file and no citations at all.

Being readable isn’t the same as being chosen

Schema and llms.txt make you legible to a machine. They don’t make you the answer. Legibility is necessary and nowhere near sufficient, which is why the entire structured-data pillar is worth 18 points and content plus third-party proof together are worth 40.

Assistants trust other people about you

When a model assembles a recommendation, it leans hardest on sources that aren’t yours — reviews, roundups, trade coverage, independent comparisons. That’s 30 points across proof and authority, and it’s the pillar most brands score worst on because it can’t be shipped in a sprint.

Comparison content is the highest-leverage page you don’t have

The prompts that produce revenue are comparative: best, versus, alternatives to. One honest comparison hub is worth more than a year of top-of-funnel blogging, which is why question one alone carries ten points.

The weighting, in full

Pillar	What it establishes	Weight
Answer-ready content	Whether there is something quotable to cite in the first place	24
Structured data & access	Whether machines can read and reach it	18
Entity strength	Whether the model is confident who you are	16
Third-party proof	Whether anyone else vouches for you	16
Authority & earned media	Whether credible publishers reference you	14
Technical & measurement	Whether crawlers see your content, and whether you’d know if things improved	12

How to read your score

0–29 · Invisible. There is effectively nothing for an assistant to cite. Recommendations in your category go to competitors by default, and no amount of technical work fixes this alone — the gap is content and proof.
30–49 · Fragmented. Some signals exist but they don’t reinforce each other. You’ll get named occasionally, usually for branded prompts where someone already knew your name.
50–69 · Competitive. The foundations are real. You appear in some category prompts but lose the high-intent comparison ones to better-proofed competitors. This is where most well-run brands land, and where the biggest gains are available.
70–84 · Strong. You’re a default candidate in most relevant prompts. Remaining points sit in earned media, original data and prompt-level measurement.
85–100 · Category authority. You’re the reference the models reach for, and competitors are working to displace you. The job shifts from building to defending.

What a low score usually costs

The commercial argument for GEO isn’t traffic volume, which is still modest compared with classic organic. It’s intent quality. A visitor who arrives because an assistant recommended you has already had the comparison done for them by a source they trust; they arrive at the bottom of the funnel rather than the top. Losing that channel doesn’t show up as a traffic drop. It shows up as competitors appearing on shortlists you were never invited to.

The second cost is compounding. Citation share is self-reinforcing — being cited generates the coverage and mentions that make future citation likelier. Brands that establish the position early are progressively harder to displace, which is the same dynamic that made early domain authority so durable in classic Google SEO.

The order to fix things in

Your fix list is already sequenced by points recoverable, but the general pattern holds across most brands:

Weeks 1–2: confirm AI crawlers aren’t blocked, verify content renders without JavaScript, publish llms.txt, fix schema. Cheap, fast, and pointless on its own — but everything downstream depends on it. This is technical SEO work.
Weeks 3–8: build the comparison and alternatives hub, restructure commercial pages to answer-first, publish pricing. This is where most of the recoverable points sit, and it overlaps heavily with AI SEO.
Months 2–4: reconcile your entity across every profile and directory, build out author entities, run a review-generation push on the platforms your category is judged on. For multi-location brands this is largely local SEO.
Months 3–12: digital PR for roundup inclusion and trade coverage, plus one piece of original research. Slowest to move, hardest for competitors to copy, and the reason scores above 80 are rare.

Questions

AI visibility, answered

Generative Engine Optimization is the practice of getting your brand named and cited inside AI-generated answers — ChatGPT, Perplexity, Gemini, Copilot and Google’s AI Overviews — rather than ranked in a list of blue links. The overlap with SEO is substantial: both reward crawlable content, topical depth and genuine authority. The divergence is where it matters. SEO optimises a page to be clicked; GEO optimises a passage to be quoted, and weights third-party corroboration far more heavily, because a model assembling a recommendation trusts what others say about you over what you say about yourself. See how our GEO services work →

No, and the difference is worth understanding. This AI visibility checker measures whether the signals that drive citation are present on your brand. A full AI visibility audit queries the assistants directly across a defined portfolio of buyer prompts and measures how often you’re actually named, in what context, and which competitor is named instead. The checker tells you why you’re losing; the audit tells you how much you’re losing. Most brands need the first before the second is actionable.

It’s a plain-text file at your domain root that summarises what your business does and points to the pages carrying your canonical answers, in a format designed for language models rather than browsers. Adoption is still uneven across AI providers, which is why it carries only four points here rather than the outsized weight some GEO checklists give it. It takes an afternoon, costs nothing and can only help — so publish it, but don’t mistake it for a strategy.

Because that’s what the models actually read when asked to recommend. Faced with “best [category] company”, an assistant doesn’t weigh up everyone’s marketing copy — it synthesises the independent sources that already made that judgement: review platforms, trade roundups, comparison articles. Your own website establishes what you do and what you charge. Third parties establish whether you’re any good. Between them, proof and authority are worth 30 of the 100 points, and they’re the hardest 30 for a competitor to take from you once you have them.

Quarterly is the right cadence. The technical and content pillars move within weeks of a focused sprint; entity, proof and authority move over quarters. Scoring monthly mostly measures noise. What’s worth doing more often is the thing question fifteen asks about — tracking your citation share across a fixed set of buyer prompts, so you can see movement in the outcome rather than only in the inputs.

Check that AI crawlers aren’t blocked, and view your page source with JavaScript disabled to confirm your content is actually there. Those two take under an hour and, when they’re broken, everything else you do is wasted effort. After that, the fastest genuinely valuable move is a single honest comparison page covering you and your three closest competitors — including the scenarios where a competitor is the better choice. Assistants reward that kind of page precisely because it reads as an assessment rather than a pitch.

Yes, and the weighting shifts in your favour. Assistants answering “near me” style prompts lean heavily on Google Business Profile data, review volume and recency, and consistent business information across directories — which are questions seven and ten here. Local businesses generally find the proof pillar easier to move than national brands do, and the earned-media pillar harder. See how local SEO and GEO combine →

It’s built by an agency that sells GEO, so the fair answer is: judge the weighting yourself, it’s published in full above. What we’ve deliberately avoided is the common trick of loading the score toward whatever’s easiest to sell as a quick project. Schema, llms.txt and technical fixes are the fastest things any agency can bill for, and together they’re worth 30 points here rather than 60. The heaviest weights sit on content and third-party proof — slower, harder, and where the actual results come from.

Keep going

Where to take your score next

The AI visibility checker names the gaps. These are the pages that explain how each one gets closed.

Generative Engine Optimization

How we build citation share inside ChatGPT, Perplexity, Gemini and AI Overviews — comparison hubs, entity work and prompt-level measurement.

Explore GEO services →

AI SEO Services

Answer-first content architecture, schema graphs and llms.txt — the work that makes your pages quotable rather than just crawlable.

Explore AI SEO →

Technical SEO Services

Crawler access, server-side rendering, Core Web Vitals and the structured data audit behind questions four to six and fourteen.

Explore technical SEO →

SEO ROI Calculator

Put a revenue number on the organic and AI-search visibility you’re currently missing, before you commit budget to closing the gap.

Open the calculator →

Local SEO Services

For multi-location brands, where Google Business Profile data and review recency decide who gets named in “near me” AI answers.

Explore local SEO →

The Paradigm blog

Working notes on AI search, generative engine optimization and the performance marketing systems we run for clients across four markets.

Read the blog →

Next Step

Find out which prompts you’re already losing

The checker tells you what’s missing. On a free 30-minute call we’ll run a live check across the buying prompts real customers use in your category — and show you exactly which competitor is being recommended in your place, and why.

Book My Free AI Visibility Call See Our GEO Services

No pitch decks · Walk away with three actionable fixes · US · UK