AI search readiness checklist
AI assistants can only recommend a business they can reach, identify and trust. This list goes in that order. Each step says how to check it yourself and what “done” looks like.
www.rrwebservices.com/checklists/ai-search-readiness-checklist/Checked November 9, 2026
AI search readiness checklist
AI assistants can only recommend a business they can reach, identify and trust. This list goes in that order. Each step says how to check it yourself and what “done” looks like.
Start here: the 8 steps that matter most
If you only have an hour or two, do these first. The full list, with how to check each one, follows.
- Allow the crawlers that put sites into AI answersNone of these is blocked from your pages in robots.txt, by name or by a “User-agent: *” rule.
- Keep your key pages indexableHomepage, service, product and key article pages have no noindex in the HTML or headers.
- Use exactly the same name, address and phone everywhereOne spelling of the name, one address format, one main phone number, on every page and profile.
- Add Organization or LocalBusiness structured data to the homepageOne JSON-LD block on the homepage with name, url, logo, and for a local business address and telephone, plus sameAs links.
- Claim and complete your main profilesProfiles claimed, with the same details as your site, a description, categories and photos.
- Make each service page answer the six questions customers askEach main service page answers all six in plain words.
- Open with a plain answer, then explainThe first sentence under each main heading answers that heading directly.
- Ask the assistants your customers’ questionsA list of 5 to 10 real questions, with who each assistant names, saved with the date.
How to use this checklist
Every step has a priority (Start here, Important, When you have time), who usually does it (You or a developer) and a rough time. Done when says what finished looks like, How to check tells you how to confirm it yourself, and Common mistake is the slip we see most. Tick the box when a step passes its check.
Steps marked with a free tool can be checked in seconds at rrwebservices.com/tools/. Sources are numbered at the end; every fact was checked against them on November 9, 2026.
Your score: ______ of 29 steps done Date: ______________
Let AI search tools in
OAI-SearchBot puts sites into ChatGPT search, PerplexityBot into Perplexity, Claude-SearchBot into Claude’s search, and Googlebot and Bingbot into Google and Copilot. OpenAI says sites that block OAI-SearchBot are not shown in ChatGPT search answers.
Done when: None of these is blocked from your pages in robots.txt, by name or by a “User-agent: *” rule.
How to check it, and the usual mistake
How to check: Open yoursite.com/robots.txt. Read every group that names one of them, and the * group.
Common mistake: Pasting a “block all AI bots” list from a blog post that includes the search crawlers, not only the training ones.
User-agent: OAI-SearchBot Allow: / User-agent: PerplexityBot Allow: / User-agent: Claude-SearchBot Allow: /
Training crawlers are a different choice from search. GPTBot (OpenAI) and ClaudeBot (Anthropic) collect training data; the Google-Extended token controls whether Google may use your content for Gemini training and grounding. Google says Google-Extended does not affect inclusion or ranking in Google Search, and OpenAI says blocking GPTBot does not remove you from ChatGPT search.
Done when: You have made a choice and robots.txt says it: allowed, or a Disallow for each training token you want to refuse.
How to check it, and the usual mistake
How to check: Look for GPTBot, ClaudeBot, Google-Extended and CCBot (Common Crawl) groups in robots.txt.
Common mistake: Blocking training crawlers and assuming it keeps you out of AI answers. It doesn’t: answers also come from search crawlers and live fetches.
# Optional: no AI training User-agent: GPTBot Disallow: / User-agent: ClaudeBot Disallow: / User-agent: Google-Extended Disallow: /
When a person asks an assistant to read a page, ChatGPT-User, Claude-User or Perplexity-User fetches it. OpenAI says robots.txt rules may not apply to ChatGPT-User, and Perplexity says Perplexity-User generally ignores robots.txt, because a person asked.
Done when: Anything private is behind a login, not just disallowed in robots.txt.
How to check it, and the usual mistake
How to check: Try opening your private pages in a private browser window. If they load, they’re public.
Common mistake: Treating robots.txt as a lock. It’s a request to crawlers, and Google says it isn’t a way to keep a page out of Google either.
Some hosts, firewalls and CDNs have one-click settings to block AI bots or challenge “unusual” traffic. They can block a crawler before it ever reads robots.txt.
Done when: Bot settings on your host or CDN match what your robots.txt says.
How to check it, and the usual mistake
How to check: Look in your CDN or firewall dashboard for bot or AI settings, and in your server logs for 403 errors to these user agents.
Common mistake: Leaving a “block AI scrapers” switch on after deciding to allow AI search.
AI answers lean heavily on what search engines have indexed. A noindex in the page or in the X-Robots-Tag header keeps a page out.
Done when: Homepage, service, product and key article pages have no noindex in the HTML or headers.
How to check it, and the usual mistake
How to check: Run the page through the SEO Health Dashboard, or use URL Inspection in Google Search Console.
Common mistake: A “discourage search engines” setting or SEO plugin toggle left on after launch.
Crawlers skip pages that return errors, redirect in loops or show a “not found” page with a 200 status (a soft 404).
Done when: Each key page answers 200 directly, or after one redirect.
How to check it, and the usual mistake
How to check: The SEO Health Dashboard shows the status and any redirects.
Common mistake: Single-page apps that show “page not found” but still answer 200.
Google renders JavaScript, but says “not all bots can run JavaScript.” Several AI crawlers read only the first HTML response.
Done when: Your headings, main text and prices appear in View Source, not only after scripts run.
How to check it, and the usual mistake
How to check: Compare the raw and rendered page with the Render Analyzer, or press Ctrl+U and search for a sentence from the page.
Common mistake: Loading service descriptions, reviews or prices from a script after the page opens.
Google says it can only find links that are <a> elements with an href. Buttons and click handlers don’t count.
Done when: Menus and in-text links are <a href="..."> to real addresses.
How to check it, and the usual mistake
How to check: Right-click a menu item: “Open link in new tab” should be offered.
Common mistake: Menus built from buttons or links with href="#" that only work with JavaScript.
A proposed standard: a Markdown file at /llms.txt that points AI tools to your most useful pages. The only required part is an H1 with your name; then a short summary in a blockquote, and H2 sections listing links. Major AI companies haven’t said they use it, so treat it as cheap insurance.
Done when: yoursite.com/llms.txt loads, starts with “# Your name”, and its links all work.
How to check it, and the usual mistake
How to check: Paste your site into the llms.txt Checker. It tests the format and every link.
Common mistake: Listing every page. It should be a short guide to the best ones, with an “Optional” section for the rest.
# Mile High Heating & Air > Furnace and AC repair in Denver, same-day service. ## Services - [Furnace repair](https://example.com/furnace-repair/): prices, signs you need it ## Optional - [Blog](https://example.com/blog/)
Make it unmistakable who you are
Assistants match mentions across your site, your Google Business Profile, directories and reviews. Different versions look like different businesses.
Done when: One spelling of the name, one address format, one main phone number, on every page and profile.
How to check it, and the usual mistake
How to check: Search your phone number in quotes on Google and look at every listing that comes up.
Common mistake: An old address on a directory, or “Ltd”, “&” and “and” used differently in different places.
It states, in a form machines read, your name, website, logo, address, phone and profiles.
Done when: One JSON-LD block on the homepage with name, url, logo, and for a local business address and telephone, plus sameAs links.
How to check it, and the usual mistake
How to check: Run the homepage through the Schema Generator’s “Already have structured data?” check, or Google’s Rich Results Test.
Common mistake: Two plugins each adding their own Organization block with different details.
{ "@context": "https://schema.org", "@type": "LocalBusiness", "name": "Mile High Heating & Air", "url": "https://example.com/", "logo": "https://example.com/logo.png", "telephone": "+1-303-555-0142", "address": {"@type": "PostalAddress", "streetAddress": "1840 Wazee St", "addressLocality": "Denver", "addressRegion": "CO", "postalCode": "80202"}, "sameAs": ["https://www.linkedin.com/company/example"] }Two-way links (site to LinkedIn, Google Business Profile, YouTube and so on, and back) show they belong to the same business.
Done when: Profiles are in the footer and in sameAs; each profile links to your website.
How to check it, and the usual mistake
How to check: Open each profile and click its website link.
Common mistake: Profiles that link to an old domain or a link-in-bio page instead of your site.
It answers “who is behind this?”: the people, how long you’ve been doing this, where you work, how to reach you.
Done when: An About page with named people, real photos, history and contact details, linked from every page.
How to check it, and the usual mistake
How to check: Read it as a stranger: could you tell who you’d be dealing with?
Common mistake: A page of slogans with no names or facts.
Articles with a named author and a short bio are easier to trust and to attribute.
Done when: Each article shows an author linked to a bio page with their experience.
How to check it, and the usual mistake
How to check: Open three recent posts and look for the byline.
Common mistake: “Admin” or “Team” as the author.
Google Business Profile for local businesses, LinkedIn for B2B, and the directories and review sites of your industry are where assistants find third-party facts about you.
Done when: Profiles claimed, with the same details as your site, a description, categories and photos.
How to check it, and the usual mistake
How to check: Ask the AI Visibility Checker your customers’ questions and look at which sites the answers cite.
Common mistake: Unclaimed duplicate listings with old phone numbers.
Answers about “best X in Y” often come from lists, reviews and directories. Being named there matters as much as your own site.
Done when: You know the 5 to 10 sites cited for your main questions and are listed or mentioned on several.
How to check it, and the usual mistake
How to check: The AI Visibility Checker shows who is named instead of you and why.
Common mistake: Buying placements on low-quality “best of” lists that assistants ignore.
Give them pages worth quoting
What is it, who is it for, what does it cost (or how is it priced), how long does it take, where do you offer it, how do I start. Pages that answer these get quoted.
Done when: Each main service page answers all six in plain words.
How to check it, and the usual mistake
How to check: The AEO Analyzer lists the questions a page should answer and which it misses.
Common mistake: Leaving out price completely. A range or “from” price is better than nothing.
AI tools quote short passages that make sense on their own. A first sentence like “Furnace repair is…, for…” is easy to lift.
Done when: The first sentence under each main heading answers that heading directly.
How to check it, and the usual mistake
How to check: Read only the first sentence of each section. Does it answer the heading?
Common mistake: Starting sections with “In today’s world…” or a story.
People ask assistants questions. Matching subheadings make the answer easy to find.
Done when: Two or more subheadings per key page are questions customers ask, answered right below.
How to check it, and the usual mistake
How to check: Compare your subheadings with the questions in your enquiries and reviews.
Common mistake: Questions nobody asks, written to fit keywords.
Two to four sentences per paragraph keeps each passage self-contained.
Done when: No paragraph longer than about 80 words on key pages.
How to check it, and the usual mistake
How to check: The SEO Health Dashboard flags long paragraphs.
Common mistake: Walls of text copied from a brochure.
Short questions and answers are the easiest parts of a page to quote. FAQPage markup still describes them, but Google stopped showing FAQ rich results in May 2026, so expect no star-like extras in Google.
Done when: Three to six real customer questions per key page, each answered in about 40 to 60 words.
How to check it, and the usual mistake
How to check: Use questions from your emails, calls and reviews.
Common mistake: Marking up questions that aren’t visible on the page.
Concrete details (price ranges, areas served, response times, years in business, certifications) give assistants something definite to repeat.
Done when: Key pages state specific, true facts instead of “best”, “leading”, “top quality”.
How to check it, and the usual mistake
How to check: Highlight every adjective on the page. Replace each with a fact or remove it.
Common mistake: Figures without a source, which hurt trust if checked.
Structured content is easier to read and to lift accurately.
Done when: Process steps are a numbered list; options and prices are a table.
How to check it, and the usual mistake
How to check: Look for long paragraphs that are really lists.
Common mistake: Tables built from images instead of HTML.
Prices, rules and “best of” content go stale. A visible updated date tells readers and tools how fresh it is.
Done when: Pages with prices or rules show a real “updated” date and are reviewed at least yearly.
How to check it, and the usual mistake
How to check: List pages with prices or legal details and their last review date.
Common mistake: Changing the date without changing the content.
Original data, case studies with real numbers and clear opinions from experience get cited because nobody else has them.
Done when: At least one page with your own data or results, with how you got them.
How to check it, and the usual mistake
How to check: Would this page exist anywhere else? If yes, it’s not original.
Common mistake: Rewriting what already ranks.
Check where you stand
The only way to know if you’re named is to ask, in ChatGPT, Perplexity, Claude, Gemini and Copilot.
Done when: A list of 5 to 10 real questions, with who each assistant names, saved with the date.
How to check it, and the usual mistake
How to check: The AI Visibility Checker asks them for you and lists who’s named instead.
Common mistake: Asking only about your brand name. Ask the questions a new customer would.
Clicks from answers show up as referrals from sites such as chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com.
Done when: A saved report or segment of these referrers, checked monthly.
How to check it, and the usual mistake
How to check: In your analytics, filter traffic sources by those domains.
Common mistake: Expecting every AI mention to bring a click. Many answers don’t link.
Server logs show whether OAI-SearchBot, PerplexityBot, Claude-SearchBot and others reach your pages, and what status they get.
Done when: You’ve seen these user agents in your logs with 200 responses.
How to check it, and the usual mistake
How to check: Ask your host for access logs, or use your CDN’s bot report.
Common mistake: Trusting the user agent alone. Anyone can fake one; the companies publish ways to verify their crawlers.
Crawlers revisit on their own schedule; OpenAI says robots.txt changes take about a day to apply to ChatGPT search.
Done when: A monthly re-check on the calendar.
How to check it, and the usual mistake
How to check: Run the Site AI Readiness Check again and compare.
Common mistake: Changing everything at once, so you can’t tell what helped.
The free Site AI Readiness Check runs many of these checks on your page in under a minute and tells you what to fix first.
- Overview of OpenAI crawlers
- Does Anthropic crawl data from the web, and how can site owners block the crawler?
- Perplexity crawlers
- List of Google’s common crawlers (Google-Extended)
- Introduction to robots.txt, Google Search Central
- Understand the JavaScript SEO basics, Google Search Central
- The /llms.txt file proposal
- Google drops FAQ rich results (May 2026), Search Engine Journal
Free to print and share. If you publish it, please credit RR Web Services and link to this page. More checklists