An AEO checklist is the list of checks that decide whether AI answer engines, such as ChatGPT search, Claude and Google’s AI Overviews, can reach, read and quote your pages. Start with access: the robots.txt rules for OAI-SearchBot and Claude-SearchBot, then the response your server actually sends them. After that come indexing, snippet controls, page structure, honest structured data and answer-first writing.
Why does an AEO checklist start with access?
Because nothing further down the list matters if the crawler gets an error instead of your page. Robots.txt permission and the server’s real response are two separate layers, and a site can pass the first while failing the second.
We learned this on an SEO Autopilot engagement with a US healthcare analytics company on WordPress behind a firewall. The firewall returned 403 to OpenAI’s and Anthropic’s search crawlers on every page. The trigger was a generic bad-bot rule matching the substring “searchbot”. Eleven other crawlers passed. The client’s monitoring tool reported all sixteen AI crawlers as accessible, because it read robots rules instead of sending requests. We wrote up the mechanism in how a firewall blocked AI crawlers.
So this checklist is ordered the way we audit: access first, then eligibility, then structure, then markup. Each check is phrased so you can answer it yes or no.
Access checks: robots.txt and the real server response
Confirm that the search crawlers are allowed in robots.txt, then confirm that your server actually serves them a page. Treat training crawlers as a separate decision.
OpenAI’s overview of its crawlers says OAI-SearchBot is used to surface websites in search results in ChatGPT’s search features, that sites opted out of it will not be shown in ChatGPT search answers, and that each setting is independent of the others, so you can allow OAI-SearchBot while disallowing GPTBot. Anthropic’s help centre article on its crawlers says Claude-SearchBot analyses content to improve the relevance and accuracy of search responses, and that disabling it may reduce your site’s visibility in user search results.
- Robots.txt has no rule that blocks OAI-SearchBot from pages you want cited in ChatGPT search.
- Robots.txt has no rule that blocks Claude-SearchBot from those same pages.
- GPTBot and ClaudeBot, the training crawlers, are allowed or disallowed as a deliberate choice, separate from the search crawlers.
- You have decided on the user-initiated agents, ChatGPT-User and Claude-User. OpenAI says robots.txt rules may not apply to ChatGPT-User because those visits are initiated by a user.
- Key pages, not only the homepage, return 200 when requested with each search crawler’s user agent. A 403, 429 or 503 is a block, whatever robots.txt says.
- Firewall, CDN and security-plugin rules contain no substring matches, such as “bot” or “searchbot”, that catch AI search crawlers by accident.
- Requests from OpenAI’s published IP ranges are allowed. OpenAI recommends allowing them alongside the robots.txt rule.
- Every subdomain has its own robots.txt with the rules you intend. Anthropic’s instructions for blocking say to repeat the rule for every subdomain, so the same applies when you mean to allow.
- After any robots.txt change, you re-test a day later. OpenAI says it can take ~24 hours for its systems to adjust for search results.
For check five, a user-agent request is a useful first signal but it is your own request, not the crawler’s, so a firewall that also looks at IP addresses may treat the real crawler differently. Our guide to testing robots.txt walks through the request-level tests.
What does Google require before its AI features can use a page?
The same things it requires for ordinary search, plus one Search Console setting. Google’s guide to optimising for generative AI features says that to be eligible for AI Overviews and AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet.
The same guide says a site must be included in Search generative AI features in Search Console to be eligible for display in them, and that from Google Search’s perspective, optimising for generative AI search is still SEO. That makes these checks the Google half of any GEO checklist:
- Each page you want cited is indexed in Google, which you can confirm with URL Inspection in Google Search Console.
- The site is included in Search generative AI features in Search Console.
- Pages you want quoted carry no nosnippet rule. Google’s featured snippets documentation says nosnippet blocks all snippets, featured and regular, and takes priority over data-nosnippet.
- The paragraph that answers the page’s main question is not wrapped in a data-nosnippet attribute.
- Every URL in your XML sitemap returns 200 rather than a redirect or an error. This is our own practice: it keeps the list you hand to crawlers clean.
One thing you cannot tick: Google says you can’t mark a page as a featured snippet. Its systems decide whether a page would make a good one. The checks above make a page eligible; they do not make it chosen.
How should an AEO checklist handle page structure?
Write each page so a single passage can answer a single question without the rest of the page around it. Answer engines quote passages, so the structure decides what can be lifted.
- The page has exactly one H1, and it names the topic.
- The first paragraph answers the page’s main question in plain words, before any background.
- H2s are phrased as the questions a reader would type, and each section opens with a one or two sentence direct answer.
- Procedures are numbered lists and comparisons are real HTML tables, not images of tables.
- A short FAQ uses questions written the way people type them, with answers that stand on their own.
- The author’s name and the publication date are visible on the page.
Check fifteen sounds trivial. On the healthcare analytics site, 21 of 39 pages had no H1 because the post template rendered the title as an H2. One template change took pages with exactly one H1 from 9 of 39 to 30. Structure problems tend to live in templates, which is good news: one fix reaches every page.
Does structured data belong on an AI SEO checklist?
Yes, as part of normal SEO, but not as a special AI requirement. Google’s generative AI guide says structured data isn’t required for generative AI search and there is no special schema.org markup to add, while recommending you keep using it because it helps with eligibility for rich results.
Google’s general structured data guidelines set the rules worth checking:
- Markup describes only content that is visible to readers of the page, with no fake reviews or unrelated content.
- Every required property for the rich result type you use is filled in. Items missing required properties are not eligible.
- Structured data sits on the page it describes, and that page is not blocked to Googlebot by robots.txt, noindex or any other access control.
Google also says using structured data enables a feature to be present but does not guarantee it will be. Treat markup as eligibility, not a lever.
- If you publish an llms.txt file, it is for services that use it. Google says Google Search ignores llms.txt and that it will neither harm nor help your visibility there.
- You re-measure against a frozen baseline: the same URLs, the same rubric, before and after each change.
What we have seen when the access layer is fixed
On that US healthcare analytics engagement, the baseline was frozen on 15 July 2026 and re-measured on 18-19 August 2026 by fetching all 39 sitemap URLs. Our AI Visibility score went from 18.35 to 66.35. Within it, AI crawler access went from 45 to 82, schema and structured data from 5 to 78, and llms.txt from 0 to 85. AI review signals did not move, 18 to 18.
It was not all gain. Performance fell from 48 to 28, because removing the firewall rule took the CDN edge cache with it, and median time-to-first-byte rose from 0.42s to 2.39s. That is why check twenty-five matters: re-measure everything, including what you were not trying to change.
The AI Visibility score weights AI schema and structured data at 25%, AI crawler access at 20%, llms.txt at 15%, AI content readability at 15%, AI content gap at 15% and AI review signals at 10%. Those weights are our rubric, frozen at baseline, not a measure of how any answer engine ranks pages.
How to run the AEO checklist on your own site
Work top to bottom and stop to fix anything in the access group before moving on, because a blocked crawler makes every later check irrelevant.
| Group | Checks | How to verify |
|---|---|---|
| Access | 1 to 9 | Read robots.txt on each host, then request key pages with each crawler’s user agent and record the status code |
| Google eligibility | 10 to 14 | Search Console URL Inspection and settings, page source for nosnippet rules, sitemap status codes |
| Page structure | 15 to 20 | Page source of one page per template |
| Structured data | 21 to 23 | Rich Results Test and a visual comparison with the visible page |
| Housekeeping | 24 to 25 | Decide on llms.txt; record a baseline before changing anything |
For a first pass, our free AI Visibility Checker reads robots.txt rules for 11 AI crawlers, reads llms.txt and the sitemap, and requests the homepage once with each crawler’s user agent. That is the checker’s own request, not the real crawler’s, and it covers the homepage only, so treat it as a first check, not proof of access. If you would rather have the whole list run and fixed every month, SEO Autopilot is ₹20,000 a month (about US$230) for one site, with no setup fee and no minimum term.
Frequently Asked Questions
What is an AEO checklist?
A list of yes-or-no checks covering whether AI answer engines can reach your pages, whether Google treats them as eligible, and whether the pages are structured so a passage can be quoted. Access checks come first.
Is a GEO checklist different from an AEO checklist?
Not in practice. Google’s guide notes the two terms stand for answer engine optimisation and generative engine optimisation, and says optimising for generative AI search is still SEO. We use one checklist for both.
Does allowing OAI-SearchBot in robots.txt mean ChatGPT can reach my site?
No. Robots.txt only states permission. A firewall or CDN rule can still return 403 to the crawler, so you also need to check the response your server actually sends.
Do I need llms.txt for AI search?
Not for Google. Google says Google Search ignores llms.txt and it neither harms nor helps visibility there. Publish one only if a service you care about uses it.
Can I block AI training but still appear in AI search?
For OpenAI, yes: its documentation says the settings are independent, so you can allow OAI-SearchBot and disallow GPTBot. Anthropic lists ClaudeBot and Claude-SearchBot as separate crawlers too.