Included in the free scan
GEO and AI visibility
We test how well a generative model can describe your business from your public footprint alone, and where the description goes wrong.
What we check
- Whether an AI assistant can state what you do, for whom, and why you
- Accuracy of the description against your intended positioning
- Presence of the corroborating third-party sources models lean on
- Whether you appear in category recommendation answers
- Content formats models reliably cite
Why it matters commercially
A growing share of research never touches your website. If the model's summary of you is generic or wrong, that summary is your brand for that buyer.
Example findings
- Model describes the brand by its industry only, with no differentiator.
- No third-party source corroborates the brand's main claim, so models omit it.
The checklist we run
8 fixed checks. Each one is graded pass, needs attention, failed or not applicable, with the fact we observed printed beside the verdict. The category score is the weighted pass rate of these checks — nothing is invented.
An assistant can describe you accurately
Site content alone supports a correct description of the business.
high impact · high effort
The entity is unambiguous
Name, location and category are stated plainly and repeatedly.
high impact · low effort
About page carries the facts
Founding, people, location and scope are on one page.
medium impact · medium effort
llms.txt published
An llms.txt file guides model access to key content.
low impact · low effort
AI crawlers are not blocked
Robots.txt permits the major AI crawlers.
high impact · low effort
Organization or Person schema
Identity schema is present with sameAs links.
medium impact · low effort
Third-party sources corroborate
Independent sources repeat the same brand facts.
high impact · high effort
Claims are citable
Distinctive, quotable statements exist for a model to cite.
medium impact · medium effort
Where the evidence comes from
Your live site, fetched directly
robots.txt directives per AI and search crawler, sitemap presence and size, llms.txt, structured data types, page titles, descriptions, headings and image alt text. No key required, so this source is always available.
AI assistant visibility probe
A fixed set of buyer-style prompts run through an AI assistant, recording whether your brand is named, whether your site is cited and which other brands appear. Directional, not a ranking.
If a source is unreachable we mark it unavailable and exclude it. A blocked source lowers data coverage, never your score.
What the deep dive adds
On any paid plan you can run this category on its own. The deep dive re-reads the pages we retrieved with only these 8 checks in front of it, returns a verdict page by page, prints the measured data behind each finding, and orders the fixes by impact and effort. It uses one audit from your monthly allowance.
What you need to provide
Nothing beyond your website address. This category is scored from public data we can retrieve ourselves.
Industry brand audits
See how this section is scored for your sector: accounting & bookkeeping firms brand audit, architecture & design studios brand audit, automotive brand audit, biotech & life sciences brand audit. Browse all industries.
Read more on this
How AI Assistants Decide Which Brands to Recommend
ChatGPT, Gemini and Perplexity don't rank pages — they recommend brands. Here's how generative engines choose which brands to name, and the signals that get yours picked.
Mapping the B2B Brand Journey From AI Answer to Signed Deal
Learn how modern B2B buyers move from AI assistant synthesis to signed contracts, and how to eliminate narrative drift across every digital touchpoint.
How to Get Your Brand Cited by ChatGPT and Perplexity
Learn how to get your brand cited by ChatGPT, Perplexity, and AI search engines through entity clarity, structured content, and technical corroboration.
The other scored categories
Brand foundation
Reputation & social
Website experience
Paid media
Local presence
Editorial policy
Questions
GEO & AI visibility: common questions
How do you test what AI assistants say about us?
We run a small fixed set of buyer-style prompts and record whether the brand is named, whether the website is cited, and which other brands appear in the same answers. The results are shown as a prompt panel with every prompt and answer summarised.
Is that a ranking?
No, and we label it directional. Assistant answers vary between runs, accounts and regions. Treat it as a sample of how your brand is described, not a guaranteed position.
What else does GEO cover?
Whether AI crawlers are permitted in robots.txt, whether an llms.txt exists, and whether the site publishes the machine-readable facts an assistant needs to describe you correctly.
Which plan includes this category?
It is scored on the free public-data scan, and again in every paid audit with the added detail your intake unlocks.
What does the deep dive add over the overview report?
The overview scores this category in one pass over your site. The deep dive re-reads the same pages against all 8 checks on their own, returns a page-by-page verdict, prints the measured data behind each finding, and orders the fixes by impact and effort. It uses one audit from your monthly allowance.
Want it done for you?
The audit tells you what to fix. If you'd rather not do the fixing, The Branded Agency can take it from here.
Audit-holder rate
Copywriting rewrite
We take the findings from your audit and rewrite the messaging across your site: one clear promise, consistent supporting messages, and copy that search and answer systems can quote correctly.
Start the rewrite
Work with us directly
Strategic brand engagement
A strategist-led engagement that fixes the brand underneath the copy: positioning, story, personality and the message architecture every channel runs on.
Book your strategy call
The report identifies the signals. We find the causes.
Diagnosis is free. Fixing the cause is what our brand strategy and marketing services are for — see the results across our client work.