What Makes a Website Visible to AI Search? A Practical Checklist

A website is visible to AI search when an AI assistant can reach it, read it without running JavaScript, and understand exactly who the business is and what it offers. In practice that comes down to eight checks: crawler access, server-rendered content, one consistent company profile, structured data, pages that answer real questions, specific facts, matching information on other sites, and ongoing monitoring. None of them requires a new kind of website. They require a website built with a second reader in mind.

This checklist explains each point, why it matters and how to check your own site in a few minutes.

How AI assistants find a business

AI assistants learn about businesses in two ways. The first is training data: text collected from the web months or years before you ask a question. You cannot update that on demand. The second is live retrieval: when you ask something current or specific, the assistant searches the web, reads a handful of pages and writes an answer from them, usually with links.

Live retrieval is where a website has real influence, and each assistant retrieves differently:

• ChatGPT search uses OpenAI's own crawler, OAI-SearchBot. OpenAI says sites that block it will not be shown in ChatGPT search answers.
• Perplexity uses PerplexityBot to build the index behind its answers, and Perplexity-User to fetch pages when someone asks a question.
• Google AI Overviews and AI Mode are part of Google Search. They draw on the same index Googlebot builds, so a page has to be indexed and eligible to show a snippet.
• Microsoft Copilot grounds its web answers in Bing search results.

So "AI search" is not one system. But every one of these paths starts with a crawler reading your pages. That is where the checklist begins.

1. Let the right crawlers in

Open yourdomain.com/robots.txt and read it. Look for Disallow rules under OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot or Bingbot, and for a blanket User-agent: * / Disallow: /.

Then check the layer in front of your website. Security services and CDNs can block AI crawlers before robots.txt is ever read. Cloudflare, for example, began blocking AI crawlers by default for new domains in 2025 and manages this in its AI Crawl Control settings. If your site sits behind a firewall or CDN, confirm the search crawlers above are allowed there too.

Search crawlers and training crawlers are separate. You can allow OAI-SearchBot (so you can appear in ChatGPT search) while disallowing GPTBot (which collects training data). We cover the details in Can AI Crawlers Read Your Website?.

2. Put the content in the HTML

Most AI crawlers read the HTML your server sends and stop there. A large analysis of crawler traffic published by Vercel found that the major AI crawlers fetch JavaScript files but do not execute them. If your services, prices or case studies only appear after JavaScript runs in the browser, those crawlers see an empty shell.

How to check: open a page, choose View Page Source (not Inspect), and search for a sentence from the middle of the page. If you can find it in the source, crawlers can read it. If the source is mostly <script> tags, the content is being built in the browser.

WordPress sites are usually fine here, because WordPress sends finished HTML. The risk is in page-builder widgets, sliders, tabs and "load more" sections that fetch content afterwards.

3. State one consistent company profile

An assistant has to decide what your business is before it can recommend it. Make that easy. Your homepage, About page and footer should state the same facts in the same words:

• the company name, exactly as you want it cited
• what you do, in one plain sentence
• who you serve and where
• how to contact you

Inconsistency is the most common problem we find: a different business description on every page, an old service still listed, a location that changed. When the facts disagree, assistants either hedge or pick the wrong version.

4. Mark it up with structured data

Structured data (schema.org markup, usually JSON-LD) restates the key facts in a format machines do not have to interpret. For most businesses, four types cover the essentials:

• Organization or a more specific type such as ProfessionalService or LocalBusiness: name, description, address, service area, contact.
• Service with Offer: what you sell and, where you publish them, prices.
• Article or BlogPosting: author and dates for articles and case studies.
• BreadcrumbList: where each page sits in the site.

Structured data does not replace clear writing on the page. It confirms it. Only mark up what the visitor can also see.

5. Answer real questions, first

AI answers are assembled from passages. A page that opens with a direct answer is easier to quote than one that opens with a slogan. For each important page and article:

• Put the answer to the page's main question in the first paragraph.
• Use headings that match how people ask ("How much does…", "What is included…").
• Add a short FAQ with questions your customers actually send you.

A useful source of questions is your own inbox: the questions prospects ask before they buy are the questions they are now asking an assistant.

6. Be specific

Vague pages give an assistant nothing to repeat. "We offer quality service" cannot be cited. These can:

• the towns or regions you serve
• the specific services you provide, by name
• prices or starting prices, if you publish them
• how the work is done, step by step
• real projects with real photos

A roofing contractor that lists named towns and explains how a roof replacement is done, as AA Home Improvements does, gives both homeowners and assistants something concrete to work with.

7. Make other sources agree

Assistants rarely rely on one website. They also read business listings, review sites, LinkedIn, industry directories and news mentions. If your Google Business Profile says one thing and your website says another, you have created doubt.

Check the profiles you control: Google Business Profile, Bing Places, LinkedIn, and the directories that matter in your industry. Make the name, description, services and contact details match your website.

8. Monitor and correct

AI answers change as models, indexes and sources change. Make a short list of questions a customer might ask, such as your company name, "best [service] in [city]" or "[your service] for [your type of customer]", and check what the main assistants say every month or quarter. Note what is wrong, trace it to its source and fix the source. We describe a repeatable method in How to Check What ChatGPT, Perplexity and Google AI Say About Your Business.

A quick self-check

What this looks like in practice

This website is built to the same checklist. Every page is rendered on the server, so crawlers receive the full text. Each page carries structured data: an Organization profile on the homepage, Service and Offer markup on the pricing page, Article markup on every case study. The company is described the same way throughout. Search and AI crawlers are allowed in robots.txt.

None of this guarantees that an assistant will recommend a business. Assistants weigh many sources, and they decide for themselves. What the checklist does is remove the reasons they can't: blocked crawlers, invisible content, conflicting facts. That is the part a website controls.