GEO AUDIT · DEEP DIVE · demo-client.com
AI-visibility audit — sample report
41 / 100 · invisible-to-partial
1 · The five signals
Crawler reachability FAIL
GPTBot and PerplexityBot receive HTTP 403 from the WAF while Googlebot passes — the site is invisible to major answer engines at the fetch layer.
robots.txt PASS
No AI crawler is disallowed; robots.txt is consistent with the WAF intent but not with its behavior (see above).
Structured data WARN
Article schema present on 12 of 34 key pages; no Organization @id graph, so entity signals don't consolidate.
Answer-first structure WARN
3 of 10 money pages open with an extractable 40–60-word answer; the rest bury the answer below the fold.
llms.txt FAIL
No llms.txt; competitors A and B both serve curated files with described links.
2 · Prompt-sampling log (excerpt — 3 of 8 prompts)
| Prompt | Named by | Note |
|---|---|---|
| best {category} software for smb | Perplexity ✗ · GPT-4o ✗ · Gemini ✓ | Named only by Gemini, position 4 of 6 |
| {brand} vs {competitor} | Perplexity ✓ · GPT-4o ✓ · Gemini ✓ | Cited via third-party review, not own site |
| is {category} worth it | Perplexity ✗ · GPT-4o ✗ · Gemini ✗ | Zero visibility — competitor C owns this set |
The full report logs all 24 answers with dates, positions, and which competitor took each slot — that's the baseline your re-check is measured against.
3 · The roadmap, staged by impact
01
Unblock the crawlers (week 1)
One WAF rule change. Verified by re-probe. Every other fix scores zero until this lands.
02
Answer capsules on 10 money pages (weeks 1–2)
40–60-word extractable answers under question-shaped H2s, per page-type template.
03
Entity graph + llms.txt (week 2)
Organization/Person @id graph sitewide; curated llms.txt with described links.
04
Off-site mentions (weeks 3–6)
The 6 sources the sampled engines already cite for this category, ranked by attainability.
4 · This page exports like the real report
Every paid report ships in three formats. Try them on this demo right now:
You are an engineer working on the website demo-client.com. Below is an AI-visibility (GEO) audit — the signals that decide whether AI answer engines (ChatGPT, Perplexity, Google AI Overviews) can crawl, understand and cite this site. Implement the fixes. ## Audit findings ### Sitewide signals - [FAIL] Crawler reachability: GPTBot and PerplexityBot receive HTTP 403 from the WAF... - [WARN] Structured data: Article schema on 12 of 34 key pages; no Organization @id graph... - [FAIL] llms.txt: absent; competitors A and B both serve curated files... ## Implementation rules 1. Fix reachability FIRST — every other fix scores zero until crawlers get HTTP 200. 2. One Organization @id graph sitewide; connect Article.publisher to it. […continues with per-fix acceptance criteria — the full version ships with every report]
You are an engineer working on the website demo-client.com. Below is an
AI-visibility (GEO) audit — the signals that decide whether AI answer
engines (ChatGPT, Perplexity, Google AI Overviews) can crawl, understand
and cite this site. Implement the fixes.
## Audit findings
### Sitewide signals
- [FAIL] Crawler reachability: GPTBot and PerplexityBot receive HTTP 403 from the WAF...
- [WARN] Structured data: Article schema on 12 of 34 key pages; no Organization @id graph...
- [FAIL] llms.txt: absent; competitors A and B both serve curated files...
## Implementation rules
1. Fix reachability FIRST — every other fix scores zero until crawlers get HTTP 200.
2. One Organization @id graph sitewide; connect Article.publisher to it.
[…continues with per-fix acceptance criteria — the full version ships with every report]5 · What you actually receive
Scored report on a private link
0–100 score, per-signal findings, per-page audit of up to 12 key pages. A human reviews every report before release; the link stays live for re-reading.
PDF + Markdown, built in
The report prints to a clean PDF and downloads as Markdown — drop it in Notion, a ticket, or forward to your agency. Try both on this very page below.
The AI fix-it prompt
A ready-to-run engineering brief: paste it into Claude Code, Cursor or ChatGPT and the agent implements the fixes — findings, rules and acceptance criteria included.
AI share-of-voice baseline (Deep Dive)
8 category prompts sampled across Perplexity, GPT-4o and Gemini — 24 logged answers showing who engines actually name in your niche.
Staged fix roadmap
Ordered by dependency, not by ease — what to fix first so everything else starts counting, with a playbook link per step.
Recorded video walkthrough (Deep Dive)
The findings and roadmap, narrated on video — async, rewatchable, shareable. No call to schedule.
Frequently asked questions
What exactly do I get for the money?
Snapshot ($49): the scored 5-signal audit plus a per-page audit of up to 12 key pages, delivered on a private link with PDF and Markdown exports and the AI fix-it prompt. Deep Dive ($99): everything in Snapshot plus the AI share-of-voice baseline (8 category prompts × 3 engines, 24 logged answers) and a recorded video walkthrough.
How fast does the report arrive?
Scope is confirmed by email within 1 business day of payment; the report lands in 2–7 business days depending on tier. Data collection is automated — the wait is the human review, not a queue.
Is it automated or does a human look at it?
Both, in that order. Probes, page audit and prompt sampling run automatically — the same engines behind the free checker. Then a human reviews every report before it's released; nothing ships sight-unseen.
Do you need access to my site or analytics?
No. Everything is measured from the outside, exactly as AI crawlers see you. Nothing to install, no accounts to share.
How is this different from the free checker?
The free check probes 5 signals. The paid audit goes page-by-page across up to 12 key pages, samples real AI answers for your category (Deep Dive), orders the fixes by dependency, and ships as a human-reviewed report with exports and the fix-it prompt.
I'm not technical — what do I do with it?
Two built-in paths: hand the PDF to whoever runs your site, or paste the AI fix-it prompt into a coding agent (Claude Code, Cursor, ChatGPT) and it implements the fixes. The roadmap tells either what to do first.
How do I pay?
Crypto checkout via NowPayments (BTC, ETH, USDT and 200+ others) directly on the audit page. Need an invoice or another method? Email hello@geoplaybooks.com.
How will I know the fixes worked?
Your public checker result page is the before/after receipt — re-run it free anytime. Prefer not to remember? Monitor re-audits every signal monthly from $20/month and emails the trend.