More customers from Google, AI and social
Sign inStart free
AI-native DeepTechDeepSEODeepAEOAI agentsGenerative searchHuman-approved

AEO checklist: 27 checks to make your website ready for AI search

By the SEOCells team, OmCells R&D Labs LLP · Published

An AEO checklist tests whether AI assistants can reach, understand and quote a website. It covers five areas: access (crawlers, indexing, speed), answers (question headings with short direct answers), evidence (facts, sources, authors, dates), structured data (organisation, page, products) and reputation (consistent details, independent mentions). Work through them in that order.

Part 1: Can AI crawlers and search engines reach the page?

  1. robots.txt allows the search crawlers you want: Googlebot, Bingbot, OAI-SearchBot, PerplexityBot, Claude-SearchBot.
  2. No firewall, bot-protection or CDN rule blocks those crawlers by mistake (check your provider's bot settings).
  3. The page returns status 200, is in the XML sitemap and has no accidental noindex.
  4. The main text is in the HTML the server sends — not only produced by JavaScript after loading.
  5. One clean address per page: a canonical URL, HTTPS, and redirects for old or duplicate versions.
  6. Pages load fast on phones: Google's Core Web Vitals “good” thresholds are LCP within 2.5 s, INP within 200 ms and CLS under 0.1.

Part 2: Does the page answer questions directly?

  1. The title and the H1 name the page's topic in the words customers use.
  2. The page opens with a 40–60 word plain answer to its main question, without links.
  3. Sub-headings are the questions customers ask, and the first sentences under each answer it.
  4. Each paragraph has one job and names its subject, so it makes sense when quoted alone.
  5. Comparisons, steps and specifications are in tables or lists, not buried in long paragraphs.
  6. The page stays on its topic; side topics have their own pages, linked from here.

Part 3: Is there evidence an assistant can trust?

  1. Key claims carry a number, a date or a named entity — and a link to a primary source where one exists.
  2. The author or reviewer is named, with relevant experience and a profile page.
  3. The published and updated dates are visible on the page and match the structured data.
  4. Reviews and ratings shown are genuine and come from customers, never invented.
  5. There is something only you can offer: your own data, photos, examples or detailed know-how.

Part 4: Is the business described in structured data?

  1. Organization (or LocalBusiness) markup with name, URL, logo, address, phone and sameAs links to official profiles.
  2. WebPage or Article markup with headline, author, datePublished and dateModified.
  3. Product or Service markup with offers (price and currency) where you sell something.
  4. Every item Google can show as a rich result has the properties Google lists as required — test it in the Rich Results Test.
  5. The markup matches what the page shows: same name, address, phone, prices and dates.

A note on FAQ markup: since August 2023 Google shows FAQ rich results only for well-known government and health websites. FAQPage markup is still valid schema.org and still describes your questions for machines, but do not expect it to change how your result looks in Google.

Part 5: Is your reputation consistent and visible?

  1. The business name, address and phone are identical on the site, in markup, on Google Business Profile and on directories.
  2. One short description of the business is used everywhere.
  3. Official social profiles link to the website, and the website links back.
  4. Independent sites — associations, press, review platforms — mention the business.
  5. You measure visits and conversions from AI assistants separately in your analytics.

Do you need an llms.txt file?

llms.txt is a proposal from September 2024 for a Markdown file at the root of a site that summarises it and lists its key pages for AI tools. It is cheap to add and can help agents that look for it, but it is a proposal, not a standard, and the major search engines have not said they use it for ranking or citing. Treat it as a helpful extra, never as a substitute for the checks above.

How can you run this checklist quickly?

The free SEOCells AEO Score runs the page-level checks above on one URL in seconds and shows what to fix first. The AEO Report adds what a single page cannot show: live AI answers sampled repeatedly, head-to-head tests against the pages that are cited, brand memory and an integrity check for hidden or misleading content.

Questions answered

What is an AEO audit?

An AEO audit checks whether AI assistants can reach, understand and quote a website: crawler access, answer-first content, evidence, structured data and reputation, followed by a prioritised list of fixes.

Does FAQ schema still help?

FAQPage markup is still valid and describes your questions for machines, but since August 2023 Google shows FAQ rich results only for well-known government and health sites. Use it where the page really is a list of questions and answers.

Is llms.txt required for AI search?

No. It is a 2024 proposal that some tools read. Crawlable pages, clear answers and structured data matter far more.

Sources

  1. Google Search Central — Introduction to robots.txt
  2. IETF RFC 9309 — Robots Exclusion Protocol (2022)
  3. web.dev — Web Vitals
  4. Google Search Central — Introduction to structured data
  5. Google Search Central Blog — Changes to HowTo and FAQ rich results (August 2023)
  6. llmstxt.org — the /llms.txt proposal (Jeremy Howard, September 2024)
  7. Google Search Central — Creating helpful, reliable, people-first content

Search engines and AI assistants change often; check the linked documentation for the latest details. Nothing in this guide is a promise of rankings, citations or results.

Run the page-level checks on your own page in seconds — the first AEO Score is free.

See what the AEO Score checks