Brand Audit

34 checks for what AI can find out about your brand

AI models learn about your brand from more than your own website. Whaily runs 34 checks across four categories, and tells you honestly what passed, what did not, and what we could not check at all. In the app this is the Your site section: the one page about your own property, where every other page is about somebody else's pages or answers.

ChatGPT
Claude
Gemini
Perplexity
DeepSeek
Grok

The problem

Most teams find these gaps only after a competitor uses them.

A missing Wikipedia page, a thin G2 profile, no reviews on Trustpilot, a social account nobody updates: none of these show up in a normal marketing dashboard. AI models read all of it when they form a picture of your brand. Most teams only notice a gap when a competitor is cited instead of them, and someone finally asks why.

How it works

From sign-up to signal in minutes.

1

Whaily runs 34 checks across four categories

Knowledge checks look at sources like Wikipedia. Review checks look at directories like G2, Capterra and Trustpilot. Social checks look at your presence on the platforms AI models read. Technical checks look at the signals on your own site. Every check gets a status and one line explaining what we found.

2

See your pass rate move, and pin what to look at

A pass rate over time chart shows whether your audit is improving or slipping. For any check, you can pin the exact URL it should look at, so a check that depends on a specific page keeps evaluating the right one as your site changes.

3

See your own pages AI has actually cited

Alongside the checks, the audit lists your own pages that AI has cited, so you can see which parts of your site are already doing the work. Your scan history shows when each check last ran and what it found that time.

What you get

Everything you need, in one place.

34 checks, four categories

Knowledge, reviews, social and technical. Every check carries a status and one line explaining what we found, so you do not have to interpret a raw score.

Pass rate over time

A trend chart shows whether your audit is improving, not just where it stands today.

Pin the URL a check should look at

For checks that depend on a specific page, pin the exact URL so the check keeps evaluating the right one as your site changes.

Your own pages AI has cited

A list of your own pages that AI answers have actually cited, so you can see which parts of your site are already earning citations.

Scan history

Every run of the audit is kept, so you can see when a check last passed, failed, or could not be checked at all.

Honest about what could not run

When a check cannot complete, for example a page it needs is unreachable, it says "not checked" instead of guessing. A gap in the data is never shown as a pass or a fail.

The audit, grouped by category

Knowledge, reviews, social and technical checks, each with a status and a plain-English reason, plus your pass rate over time.

The website technical category of the brand audit, as seven check cards: llms.txt and Sitemap failing, Schema.org Organisation markup missing, and AI crawlers allowed, Meta description, Open Graph tags and HTTPS passing.

Why it matters

Your website is only part of what AI reads about you.

AI models build their picture of a brand from more than the brand's own site. Knowledge sources, review directories, social platforms and the technical signals your site exposes all feed into what a model says, and does not say, when someone asks about you. A gap in any one of these is invisible on a normal marketing dashboard, because none of it lives in your own analytics.

Whaily runs 34 checks across those four categories: knowledge, reviews, social and technical. Every check comes back with a status and one line explaining what we found, so someone without an SEO background can read it and know what to do. A pass rate over time chart shows whether the audit is improving, and you can pin the exact URL a check should evaluate so it keeps looking at the right page as your site changes.

Honesty matters more here than almost anywhere else in the product. Some checks depend on a lookup that can fail for reasons that have nothing to do with your brand: a directory that is temporarily unreachable, a page that returns an error. When that happens, the check says "not checked" instead of quietly counting as a pass or a fail. A missing answer is not the same as a bad answer, and we do not blur the two.

One check, explained

Open any check to see why it has that status, pin the URL it should look at, and see when it last ran.

One audit check opened: llms.txt marked absent and when it was last checked, why it matters, a suggested llms.txt to copy, and six numbered steps to fix it.

Questions

The short answers.

What are the 34 checks?+
They fall into four categories: knowledge sources like Wikipedia, review directories like G2, Capterra and Trustpilot, social presence, and technical signals on your own site. Each check has a status and one line explaining what we found.
What does "not checked" mean?+
It means the lookup for that check could not run, for example because a page or a third-party source was unreachable. We show "not checked" instead of guessing at a pass or a fail, so you always know the difference between a real gap and a check that simply did not complete.
Can I control which URL a check looks at?+
Yes. For checks that depend on a specific page, you can pin the URL, so the check keeps evaluating that page even if your site's structure changes.
Does the audit show which of my pages AI has already cited?+
Yes. Alongside the checks, you get a list of your own pages that AI answers have actually cited, so you can see what is already working.
Does Whaily check whether AI crawlers can access my site?+
Not yet. A check for crawler access, whether GPTBot, ClaudeBot and similar crawlers can fetch your key pages, is planned but not shipped. We will say so plainly on this page when it ships.
How often does the audit run?+
The audit re-runs on a schedule, and every run is kept in your scan history, so you can see when a check last passed, failed, or came back not checked.

Ready to be recommended by AI?

Start free. See your first insights in minutes.