Free · 60 s Free AI-readiness audit See your site the way ChatGPT does.
Personal reply within one business day.

Found · The paid starting point

AI Visibility Deep Audit

In seven business days you’ll know why ChatGPT, Perplexity, Gemini, Claude and Google’s AI Overviews name your competitors instead of you, and exactly what it costs to fix, in fixed quotes.

€1,500 · $1,650 7 business days

Fixed price. 100% credited to any build within 60 days.

Proof on file

My own previous site

0

words of page content visible to non-JavaScript crawlers

The first site I audited this way was my own. It looked finished in every browser. Crawlers received an empty shell. This site is the rebuild.

observed 2026-10-05 · raw HTML fetch

03 Fit

Who this is for, and who it isn't.

Built for you if

  • Competitors show up in ChatGPT or Perplexity answers for questions you should own, and you don’t.
  • You want a diagnosis with evidence before committing budget to a rebuild, a sprint or a retainer.
  • You need to explain AI search to partners or a board with dated findings, not opinions.
  • You changed CDN, hosting or security settings recently and something feels off.

Not for you if

  • You only want a quick score. The free AI-readiness audit gives you that in about a minute.
  • You expect a guaranteed ranking or citation afterwards. Nobody can honestly sell that, and I won’t.
  • You already know what’s broken. Go straight to the AI Search Visibility Sprint; it opens with its own baseline.
  • You sell to consumers with no considered purchase. My prompt sets are built around how business buyers shortlist.

04 The problem

You’re losing shortlists you never knew existed.

A buyer asks an assistant for the best firms for their problem. It answers in seconds, with names. If yours isn’t one of them, you never find out: no visit, no bounce, nothing in your analytics. G2 found that 51% of B2B software buyers start research in an AI chatbot, and 6sense reports that 85–95% of B2B buyers buy from their day-one shortlist.

The causes are usually ordinary and fixable: a CDN rule that quietly blocks GPTBot, pages that show nothing without JavaScript, a company description that differs on every profile, service pages that never answer a question directly. The trouble is that all of them are invisible from the inside. In your own browser, the site looks fine.

The audit looks from the outside, the way an assistant does, and puts a date and a method next to every finding.

Signs you need this

  • A competitor shows up in ChatGPT’s answers. You don’t, and you know you’re better.
  • Your SEO agency says everything is fine, but Perplexity has never heard of you.
  • Somebody changed the Cloudflare or firewall settings and nobody checked what it blocked.
  • You’ve read every “AI SEO” checklist going and can’t tell which items matter for you.
  • Assistants describe your company with an outdated service list or the wrong city.
  • Leads mention “I asked ChatGPT” and you can’t tell what it told them.

Want a first look before you pay? Run the free scan.Bot access and readability in about 60 seconds. The Deep Audit goes much deeper.

06 What you get

Deliverables, tagged by the layer they fix.

One domain, one language, your main competitors. Everything is dated and reproducible, so you can re-run it yourself next quarter.

  • AI-crawler access test, 10+ bots

    GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended and more, tested against robots.txt, Cloudflare bot rules and firewall traps that block them silently.

    Layer: Found
  • No-JavaScript render test

    What a crawler actually receives from your key pages: words, headings, links and structured data, page by page.

    Layer: Found
  • 30 buyer prompts, five assistants

    Real questions your buyers ask, run on ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews. Who is named, who is cited, and from which sources.

    Layer: Cited
  • Competitor share of voice

    How often you and the competitors you choose appear, and which pages and profiles assistants cite for them.

    Layer: Cited
  • Entity and schema review

    Is your company described the same way everywhere? Organization, Service and Person markup, sameAs links, directory and review profiles.

    Layer: Cited
  • Conversion-path and journey map

    What happens when an assistant does send someone: landing page, speed, proof, prices, form, reply time, and where they drop.

    Layer: Chosen
  • 45-minute recorded walkthrough

    I walk you through the findings on a call, recorded so partners who weren’t there can watch it later.

    Layer: All layers
  • 90-day plan with fixed quotes

    Every fix ranked by effort and impact, with a fixed price for anything I would do. Do it yourself, with your agency, or with me.

    Layer: All layers

07 How it runs

Seven business days, three human gates.

Agents run the repetitive tests at scale. I design the prompts, read every result and write every conclusion. Nothing reaches you that I can’t reproduce.

Agents at work Human gate: nothing moves on until a person approves

  1. Step 1: Day 1

    Kick-off and prompt set

    A 20-minute call. You name your competitors and the questions buyers ask before they call you. I turn them into 30 test prompts.

    • agent://researcher

    Human gate You approve the prompt set and the competitor list.

  2. Step 2: Days 1–3

    Crawl and access tests

    Crawler access, no-JavaScript render, schema extraction and speed checks across your key pages.

    • agent://researcher
    • agent://tester
  3. Step 3: Days 2–4

    Prompt runs

    Every prompt run on five assistants, with answers and citations captured and dated so they can be compared later.

    • agent://researcher
  4. Step 4: Days 4–5

    Analysis

    Findings cross-checked against the raw captures. Anything that can’t be reproduced is dropped, not reported.

    • agent://fact-checker

    Human gate I review every finding before it reaches you.

  5. Step 5: Days 6–7

    Walkthrough and plan

    The recorded walkthrough, the written report and the 90-day plan with fixed quotes.

    Human gate You decide what happens next. There is no obligation.

08 Proof

I audit with the same tools I run on my own systems.

The law-firm engine counts AI crawlers by name every day and raises an alert when they stop visiting. The audit uses the same thinking on your site, once, in depth.

CLIENT · NAME CHANGED

16 AI crawlers tracked by name

Corvane Legal: AI content engine

A German law firm’s knowledge engine with 1,100+ pages live on its own domain, per-bot crawler analytics and an alert if AI crawlers stop coming.

  • Content engine
  • Crawler analytics
CLIENT · NAME CHANGED · PRE-LAUNCH

15 would-be 404s caught

Orvelle Dental: pre-launch audit

A hop-by-hop audit of every sitemap URL, internal link and archived address, run before the new site went anywhere near live.

  • Redirect audit
  • Rebuild
0✓ words of page content visible to non-JavaScript crawlers observed 2026-10-05 · raw HTML fetch
2✓ URLs in the sitemap observed 2026-10-05 · sitemap.xml
16✓ AI crawlers tracked by name observed 2026-10-05 · crawler module

Want to see the format first? Run the free audit on your own site, then ask me for the sample Deep Audit on a call.

09 Price

What moves the price, and what it costs to run.

The audit is a fixed €1,500 · $1,650. 100% credited to any build within 60 days. Scope only changes if you ask for more, and then it’s quoted before work starts.

Quoted separately

  • A second language or market (prompts and profiles differ per market).
  • Additional domains or sub-brands.
  • A larger prompt set for complex product ranges.
  • An on-site workshop instead of the recorded call.

Always included

  • One domain, one language, your main competitors.
  • 30 buyer prompts across five assistants.
  • Recorded walkthrough and written report.
  • Fixed quotes for every fix, credited if you build with me.

Monthly running costs

Paid by you, directly to each provider, in your own accounts. I add no markup and no hosting fee.

WhatHow it worksPaid to
During the auditNothing. I use my own tools and accounts; you only grant read-only access if you want firewall rules checked.Nobody
After the auditEach fix in the 90-day plan states its own running cost, if it has one. Most have none.Stated per fix

100% credited to any build within 60 days. If you build elsewhere, the plan and quotes are yours to take with you.

10 Handover

What you own when we're done.

  • Written report with every finding dated PDF + Markdown
  • Recorded 45-minute walkthrough shared drive
  • Raw prompt captures from all five assistants spreadsheet
  • Your prompt set, ready to re-run each quarter spreadsheet
  • Crawler and firewall findings with exact rule changes report appendix
  • 90-day plan with fixed quotes PDF

No licence fees to me, no lock-in, no proprietary platform. Everything is dated and reproducible: rerun the prompt set yourself or hand the report to any agency.

11 Questions

Straight answers to fair questions.

How is this different from the free audit?

The free audit is an automated scan that takes about a minute: bot access, readability and your top fixes. The Deep Audit is seven business days of my time plus agents: 30 prompts across five assistants, competitors, entity review, the conversion path, a recorded walkthrough and fixed quotes.

Can you guarantee we’ll be recommended by ChatGPT afterwards?

No. Nobody controls what an assistant says, and anyone who guarantees it is guessing. What I can do is remove everything that stops it from finding, reading and trusting you, and measure before and after with the same prompts.

Is the fee really credited?

Yes: 100% credited to any build within 60 days. If you don’t build with me, the plan and quotes are still yours to use with anyone.

What do you need from us?

A 20-minute kick-off, the names of the competitors you lose to, and read-only access to Google Search Console if you have it. No passwords to your website.

Do you need access to our CMS or hosting?

No. Everything is tested from the outside, the way an AI crawler sees you. If you want Cloudflare or firewall rules checked directly, a read-only account is enough.

Which assistants do you test, and won’t the answers change?

ChatGPT, Perplexity, Gemini, Claude and Google’s AI Overviews. Yes, answers change, which is why every capture is dated and the prompt set is yours to re-run. The point is the pattern across 30 prompts, not one lucky answer.

Not sure the audit is the right first step?

20 minutes with me, not a sales rep. You leave with a clear next step — whether or not we work together.