GEO Readiness Scorecard: Is Your Brand Visible to AI Search?
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 scorecard 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
Answer-Ready Content
24 pointsDo 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 pointsHow 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 pointsIs 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 pointsWhat 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 pointsHave 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 pointsDoes 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.
0 of 15 answered
Your priority fixes
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 readiness. Treat the score as a diagnosis of your foundations and the fix list as the order to work in.
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 search.
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.
GEO readiness, answered
Where to take your score next
The scorecard 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 →Find out which prompts you’re already losing
The scorecard 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.
No pitch decks · Walk away with three actionable fixes · US · UK · UAE · India
