The AI search toolkit checks how ready a website is to be found and quoted by AI search, and whether it’s actually appearing. It covers AI crawler access, how quotable the content is, how clearly the business is defined, and visibility in Google’s AI Overviews and AI Mode and in assistants such as ChatGPT and Perplexity.
Built in-house. It is my own tooling rather than a reseller’s dashboard, which is why it checks the things I act on rather than everything a generic tool can measure.
What it does
The AI search toolkit runs two kinds of checks. Readiness checks look at the website: whether AI crawlers can reach it, whether key pages answer questions clearly enough to quote, whether expertise is visible, and whether structured data and business details are consistent. Visibility checks look at the answers: whether the business is mentioned or cited for a set of prompts and searches, and who’s cited instead.
The approach it follows
The toolkit follows Google’s own guidance rather than unconfirmed tricks. Google’s guide to optimising for generative AI features says its AI features in Search are rooted in its core ranking and quality systems, and that optimising for generative AI search is still SEO. It puts unique, valuable, non-commodity content first, such as first-hand perspectives rather than summaries of what’s already available. Its documentation on AI features confirms there are no additional requirements to appear in AI Overviews or AI Mode.
For other assistants, the toolkit checks access against each platform’s own documentation, such as OpenAI’s guidance on its crawlers, which explains that OAI-SearchBot surfaces sites in ChatGPT search separately from GPTBot, which relates to model training.
Readiness checks
| Check | What it looks at |
|---|---|
| Crawler access | robots.txt and firewall rules for Googlebot, OAI-SearchBot, PerplexityBot and other relevant crawlers, plus snippet controls such as nosnippet |
| Rendering | Whether key content and headings appear in the HTML, or only after JavaScript runs |
| Citability | Whether sections open with a clear, self-contained answer, and whether facts are specific and attributed |
| Structure | Heading hierarchy, FAQs and whether the page is easy to navigate |
| Expertise | Named authors, first-hand examples and visible sources |
| Entity clarity | How clearly the business, its services, location and people are defined, in content and structured data |
Visibility checks
A fixed set of searches and prompts is run on Google’s UK results and on the main AI assistants. Each is run more than once, because AI answers vary. For each, the toolkit records whether an AI answer appears, whether the business is mentioned, whether it’s linked, how it’s described, and which other businesses or sources appear. Results are dated so they can be compared over time.
Example output
SEARCH: best garage door threshold seal (google.co.uk, 15 September 2026) AI OVERVIEW: Shown. Names GaraDry’s kit first and cites GaraDry. ORGANIC: uk.garadry.com ranks first. READINESS NOTE: Product guide opens with a clear recommendation and comparison table, which is the section the answer draws on.
What I review
AI answers change often, so I never treat one result as proof. I rerun prompts that look unusual, check whether a missing citation comes from access, content or reputation, and confirm any access problem in the site’s robots.txt and server settings before reporting it. Recommendations are ordered by what’s likely to make a difference for the business, not by what’s easiest to measure.
I also don’t rely on widely shared AI search statistics unless they come with a clear method, sample and date. Your own data and the platforms’ own documentation carry more weight.
Limitations
- Answers vary. The same prompt can return a different answer minutes later, and by location and model version.
- A sample, not a census. The toolkit checks chosen prompts, not every conversation people have with an assistant.
- Readiness isn’t a guarantee. A fully ready page can still not be cited, because platforms choose sources for reasons they don’t fully publish.
Services it powers
The toolkit powers Google AI Overviews and AI Mode, ChatGPT and AI assistant visibility, AI citation readiness and AI visibility tracking. The live GaraDry result is covered in the GaraDry case study. See the rest of the tools on the tech stack page.
FAQ
What does the AI search toolkit check?
The AI search toolkit checks whether AI crawlers can reach your site, how quotable your content is, how clearly your business is defined, and whether you appear or get cited in Google’s AI Overviews and AI assistants for the searches that matter. Every result is reviewed before it’s reported.
Why don’t you use AI visibility statistics from industry studies?
Many widely shared AI search statistics come without a clear method, sample or date, or are quickly out of date. I only use figures I can trace to a primary source or to a study with its method stated, and rely on your own data wherever possible.
Does Google have an official guide to optimising for AI?
Yes. Google’s guide to optimising for generative AI features and its documentation on AI features both say foundational SEO applies to AI Overviews and AI Mode, with no additional requirements. The toolkit follows that guidance rather than chasing unconfirmed tricks such as llms.txt for Google.