Can AI Search Find Your Website? A 2026 Checklist

Photo: Samsung Galaxy S4 by Janitors, CC BY 2.0. Cropped.

AI search changed the game

For two decades "being found online" meant ranking in Google's blue links. In 2026 a growing share of answers come from answer engines — ChatGPT, Gemini, Perplexity and Google AI Overviews — that generate a response by reading and citing sources. If those systems cannot find or understand your site, you are invisible to a fast-growing slice of buyers, and no amount of traditional SEO fixes it on its own.

Three things AI needs from your site

• Findability — a clean robots.txt that lets AI crawlers in, plus an llms.txt that tells them what your business actually does.
• Understandability — clear, query-shaped content with structured data (schema) so the model knows what each page is about.
• Citeability — specific statistics, direct quotations and inline source links. Research from Princeton (Aggarwal et al.) shows these are the top signals that increase how often AI cites a page.

How to check it in 2 minutes

Paste your URL into the free Kavenacc AEO / SEO Auditor. It scores 13 dimensions — machine-readable context, AI-crawler robots.txt, schema, query-shaped content, earned media, E-E-A-T and GEO citation signals — and returns a prioritized list of what to fix first. The score, the dimension breakdown and the loophole list are free; no signup.

What a low score actually means

A low score is not a verdict — it is a roadmap. Most sites lose points on three cheap fixes: no llms.txt, a robots.txt that blocks AI crawlers, and content with no statistics or citations. Each is a same-day change. The auditor tells you exactly which, so you are not guessing.

Then make AI cite you

Once the basics are in, the highest-leverage move is adding real proof: "we shipped 50+ apps", "26 years in IT", "saved clients 15 hours/week". Numbers and named sources are what answer engines quote. Kavenacc builds AEO-ready sites, llms.txt files and AI-crawler robots.txt as standard — see the OpsCommand packages.

The 10-minute manual check

1. Open yoursite.com/robots.txt. Make sure there is no Disallow: / under User-agent: *, and that GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Google-Extended are not blocked unless you chose that deliberately.
2. View the page source (Ctrl+U in Chrome) of your homepage. Search for a sentence you can see on the page. If it isn't in the source, your content is rendered by JavaScript, and many AI crawlers will miss it.
3. Visit yoursite.com/llms.txt. If it returns a 404, you don't have one yet. See our llms.txt guide.
4. Paste your URL into Google's Rich Results Test to see whether structured data such as Organization, LocalBusiness or FAQ is detected.
5. Run PageSpeed Insights on mobile. Slow, heavy pages are fetched less reliably and convert fewer visitors.
6. Search your business name on Google. Check that your Google Business Profile, website and social profiles show the same name, phone number and services.

Why JavaScript-only websites are invisible to many AI tools

Many modern sites, including some built with React or website builders, send the browser an almost empty HTML file and then build the page with JavaScript. Google can usually render this, though sometimes with a delay. Many AI crawlers fetch the raw HTML and do not run scripts, so they see a blank page. The fix is server-side rendering, pre-rendering or a static build, so the text is already in the HTML. When Kavenacc rebuilt its own homepage this way, Lighthouse scores rose to 98–100 on mobile and AI crawlers could read every section.

Testing the answers themselves

Technical checks show whether AI can read your site. Prompt testing shows whether it chooses you. Ask each assistant five to ten questions a real customer would ask, for example:

• "Who are the best web designers in Nairobi for a small business?"
• "How much does a website cost in Kenya?"
• "Which companies build M-Pesa e-commerce websites?"

Record whether your business appears, whether the facts are correct, and which sites are cited instead. Those cited sites show you where to earn mentions: directories, industry associations, news coverage or partner websites.

What to fix first

The free AEO / SEO Auditor runs most of these checks automatically. If your site needs rebuilding to be readable by AI, a fast, AI-ready website can be launched in 7 days.

How AI search tools actually find businesses

There are two routes by which an AI assistant can learn about your business, and they need different fixes.

• Training data. Large language models are trained on a huge snapshot of the public web. If your business was well described online when that snapshot was taken, the model may "know" you without searching. You can't control this directly, but consistent, long-standing information across many reputable sites makes it more likely.
• Live retrieval. When ChatGPT search, Perplexity, Copilot or Google's AI Overviews answer a question, they often run a web search, fetch a few pages and summarise them. This is where technical access matters most: if their crawler can't fetch your page, or finds no readable text, you can't be cited, however good your business is.

For a small or newer business, live retrieval is the realistic opportunity. It rewards the same things that help customers: clear pages, fast loading, accurate facts and direct answers.

The crawler names to know

A business that wants to be recommended by AI tools generally allows all of these. If you have reasons to opt out of model training while staying visible in search answers, you can allow the search crawlers and block only the training ones, but check each provider's current documentation, as these names and their roles change.

A real example: from invisible to readable

Kavenacc's own previous homepage was a React single-page app. In a browser it looked complete, but the HTML a crawler received was essentially one empty <div>. Rebuilding it as a static page with the same design put every heading, price and FAQ answer directly in the HTML, added structured data for the organisation, founder and FAQs, and cut the page weight so that Lighthouse now scores it 98–100 across performance, accessibility, best practices and SEO. Nothing about the business changed; it simply became readable.

Frequently overlooked problems

• Cookie walls and pop-ups that cover the content before any interaction.
• Prices and services shown only in images or PDF brochures, which are harder to extract than page text.
• A "coming soon" or maintenance page left in place by a website builder.
• Duplicate sites on www and non-www, or http and https, without redirects, which splits your signals.
• Old addresses and phone numbers on directories, which AI tools may repeat as current.

Fixing these is rarely expensive, and the same fixes make the site better for human visitors too.

More questions answered

How often should we re-check?

Run the technical checks after every major website change, and test AI answers monthly. AI tools update their indexes and models regularly, so visibility can change even when your site doesn't.

Can a small business compete with big brands in AI answers?

Yes, especially for local and specific questions. A clinic in Thika or a web designer in Nairobi can be the most relevant answer for a local question, provided its pages clearly state the service, location and price, and the facts match across the web.

Related articles

AI Agent vs AI Chatbot: what is the difference and which do you need?What Is an AI Operations Command Center? (Plain-English Definition)Custom AI Software vs SaaS: Which Actually Saves You Money in 2026?

Frequently asked questions

Try the free AEO / SEO Auditor →

See Kavenacc packages & pricing →