Redacted sample

What $750 actually gets you.

A real audit with the client's name and category removed. Nothing else has been tidied up.

1. The prompt sheet

45 buying questions. For each one: named or not, position, and which source the engine cited.

PromptEngineYouNamed insteadCited source
best [redacted] for [redacted]ChatGPTnot named3 competitorsroundup blog
[redacted] vs [redacted]Perplexity#21 competitormarketplace listing
is [redacted] worth itAI Overviewsnot named2 competitorsforum thread
cheapest [redacted] that actually worksGemini#44 competitorsown site

41 more rows in the full document.

2. Your Rented vs Owned Index

The same prompts run again with retrieval disabled. This client held almost nothing.

ChatGPTnamed 9 · own 1 indicative
Perplexitynamed 23 · own 3 measured
AI Overviewsnamed 12 · own 0 indicative
Gemininamed 9 · own 0 indicative
Claudenamed 4 · own 0 measured

Read: 57 namings across five engines, 4 of them owned. Ninety-three percent of this brand's AI visibility disappears the moment an engine stops running a live search. That is not a stable position, and no dashboard on the market would have told them.

3. Machine-readability score

Itemised, with the exact defect and where it lives.

Entity resolution
Brand resolves to a different company in two engines
2 / 10 verified
Structured data
Organization present but no sameAs, no logo; Product invalid on 14 pages
3 / 10 verified
Answer-shaped content
No page directly answers 31 of the 45 prompts
2 / 10
Feed health
Clean. 0 invalid items
9 / 10 verified
Crawlability for AI agents
Two major crawlers blocked in robots.txt, probably by accident
4 / 10 verified

Every line carries how we got it. Verified means you can re-run it yourself today and get the same answer. A composite inherits the lowest confidence of its inputs, and we never average across engines.

4. What we would fix, and what we would not

Fixable by us

  • Entity disambiguation from the company they get confused with
  • Full structured data rebuild across 14 product pages
  • Answer pages for the 31 unanswered prompts
  • Unblock the two crawlers

Not fixable by us

  • The three roundup blogs that name their competitors need to name them too. That is a relationship, not a build
  • One competitor is named because it has been reviewed in a national title. We cannot manufacture that
  • Their marketplace listing outranks their own store on two prompts. That is a merchandising decision on their side

This section is why the audit is worth buying on its own. Roughly a third of what is wrong with most brands is not something we can sell them a fix for, and we would rather write that down than discover it in month three.

Get one for your category.

$750 for the first 50. Five working days. The fee comes off the build.

The fee comes off the build.

PricingFAQStart the audit

Read the full 62-section audit

The standard format, head by head — 7 parts, 35 sections, each one saying what it means, what goes in it, why it earns its place and a worked example. No client's data: the examples are invented. Download as PDF from the icon in the top-right of that page.

Open the full format — all 35 heads →