Robots & crawlers
Whether the five citation crawlers — OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, Applebot — are allowed to fetch your content. We check 27 AI crawlers in total, citation kept separate from training.
Free · no account needed
See your AI search score in 60 seconds.
ChatGPT, Perplexity, Claude and Google AI Overviews now answer the queries your buyers used to type into Google. We grade 39 signals across robots, llms.txt, schema, entity signals and AI discoverability — and email you a 0-100 score with the three fixes that move the needle. No call required, no credit card.
Real cases
What GEO work changes: before → after.
Crypto licensing · Gofaizen & Sherle
Before
Top-10 rankings were there, but the click was gone — Google started answering itself.
We did
An entity layer in place of a two-node schema: Article, Person, Service — plus a structural fix on the language version.
After
Cited in the Google AI Overview body; source #1 on the El Salvador query. First firm listed in ChatGPT.
Education · AP University
Before
A young brand in a crowded niche — no rankings, no mentions in AI answers.
We did
GEO and classic SEO as one loop: entity content, structured data, work aimed at being quotable.
After
Cited in the body of Google AI Overviews. Over 16 months: 242K impressions, 3.65K clicks, average position 12.1.
Compliance · AML certification
Before
The site was never cited in AI answers on its own industry queries.
We did
GEO plus structured data plus entity content built for LLM citability.
After
In the AI Overview body and among Perplexity sources. Over 16 months: 8.53K clicks, 1.6M impressions.
Landscaping · GardenPark
Before
A GEO audit scored 39/100 — the brand never showed up in AI answers.
We did
Audit → prioritised fixes → implementation: schema, answer-first structure, AI bot access, entity content.
After
The brand sits in the body of Google AI Overviews and ChatGPT answers for its main commercial query.
Want the same? The free score is a starting point. From there our GEO team works the brand into AI answers on your commercial queries.
All cases GEO agency Digital Delight →What is in the report
Eight categories. One score. Three fixes that matter.
The report grades you on 39 signals across the eight buckets below, then ranks every issue by impact-vs-effort. You get the full breakdown — but the executive summary names just three fixes, because that is what most teams will ship in the next sprint anyway.
llms.txt
Whether an llms.txt exists and is well-formed. Google Search ignores the file; some AI developer tools read it.
Schema markup
JSON-LD coverage across Organization, Article, Product, FAQ, sameAs and speakable.
Metadata
Titles, descriptions, canonicals, OG tags — basics that still trip half the sites we scan.
Content depth
Whether passages are extractable, answers come first, and Quick Facts blocks exist.
Brand & entity
Whether AI engines can connect your brand to a real entity — Wikidata, sameAs, named experts.
Authority signals
External citations, HARO mentions, structured author bios, E-E-A-T markers.
AI discoverability
Snapshot of how often you already appear in ChatGPT, Perplexity and AI Overviews.
What also runs, without feeding the score: whether your pages break cleanly into RAG chunks, whether comments or reviews have opened up a prompt-injection surface, where you state numbers without a source (AI engines are reluctant to quote those), which search intents your page covers and which it misses, whether your content only appears after JavaScript, and whether your CDN quietly blocks individual AI crawlers. These land in the report as findings, not as lost points.
How it works
Three steps. Most of them ours.
- 1
You submit a URL
Your site, a competitor, or both. The form takes 20 seconds. We do not ask for a phone number, a budget range, or a "team size" dropdown.
- 2
We grade 39 signals
Robots.txt and access for 27 AI crawlers, llms.txt validation, schema coverage, entity signals, content extractability, AI engine sampling. Server-side, 60-90 seconds.
- 3
You get the report
HTML report by email — score, eight category bars, three priority fixes. No call required to read it. Reply if you want a human walk-through.
What teams do with the score
Three patterns we see in the data.
62 → 88
Crypto licensing site, 8 weeks
Score jumped after fixing four items the audit flagged: GPTBot allowed, llms.txt published, FAQ schema across 40 pages, named regulatory expert with sameAs. First Perplexity citation in week three.
47 → 74
B2B SaaS, 12 weeks
Audit named one structural fix — moving the answer above the fold on every product page — and a content fix: turning the FAQ into Quick Facts blocks. ChatGPT started quoting them in the second month.
71 → 71
Local services, no change
Score stayed flat because the bottleneck was off-site, not on-site: zero entity signals, no Wikidata, no industry citations. The fix was HARO and trade press, not code — and that takes longer than 12 weeks to show.
We have run digital marketing for clients since 2020 and the AI-audit playbook since 2023. The free audit and the paid audit grade the same 39 signals — what you pay for is the human reading, the query universe research, and the engineer who will fix it.
FAQ
Honest answers about a free tool.
Is it actually free? What is the catch?
Yes — fully free, no credit card, no scheduled call. The catch is honest: if your score is below 60 and you want help fixing it, we will pitch one of our retainers. If it is above 80, we will tell you that too and probably not pitch anything.
How long does it take?
The scan runs in 60-90 seconds. The email with your full report lands in 2-5 minutes. If it has not arrived in 10 minutes, check spam, then ping [email protected].
What do you do with my URL and email?
URL goes into the audit queue and stays in our database for benchmarking (anonymous, aggregated). Email gets the report and a short follow-up sequence you can unsubscribe from in one click. We never sell or share contact data.
Can I audit a competitor site?
You can. Public URL, public score — no authentication, nothing private. Many teams run their top three competitors first to set a benchmark before scoring themselves.
How is this different from a regular SEO audit?
Traditional SEO audits grade you for Google blue links. This one grades you for being the source AI engines quote when answering a question. Different signals, different fixes — robots.txt rules for GPTBot, llms.txt presence, schema for AI extraction, entity authority. We have run both kinds professionally since 2020; this tool only does the AI side.
Why should I trust the score?
The 39 graded signals come from the open-source GEO Optimizer toolkit (MIT license, Auriti Labs) plus our own playbook layered on top. We benchmark against 500+ sites we have audited since 2023. The score is reproducible — run it twice and you get the same number.
Ready for your score?
Same form, same 60 seconds, same report.
Run my free AI auditAlready audited and want a human to read it with you? Book a free 30-minute review — no SDR layer, no obligation.