How to Do an AEO Audit: The 25-Point Checklist (2026)

Quick answer: An AEO audit checks whether AI answer engines can crawl, extract, trust, and cite your content. Run these 25 checks across five sections (retrievability, answerability, structure, authority, measurement), score one point each, and map your total to the bands below. Under 16 means AI assistants likely cannot reliably quote you, and each section links to a free tool that automates the fix.

Answer Engine Optimization is the discipline of getting AI assistants to name and cite your brand in their answers. An AEO audit is how you find out whether your site is actually built for that, or whether ChatGPT and Gemini are skipping past you to cite a competitor. This is a runnable audit. Work through all 25 checks, mark each pass or fail, add up your score, and use the section links to fix what fails.

The checklist is organized the way AI assistants actually process a page, in order of dependency. If a bot cannot retrieve your content, answerability does not matter. If your content is retrievable but unanswerable, structure does not matter. Work top to bottom.

How to Use This Checklist

Copy the checklist below into a doc or spreadsheet. For each of the 25 checks, mark it done only if it fully passes. Score one point per pass, zero per fail or partial. Total your score out of 25 and read the band. Then fix the failures section by section, starting with retrievability, and re-run the section after each round of fixes.

For an automated pass across all five sections, run the SEO audit and let it surface the failures and priorities for you. The manual checklist below is the copy-paste version you can hand to a client or a junior marketer.

Section 1: Retrievability (5 checks)

If AI crawlers cannot fetch your pages, you are invisible to answer engines no matter how good the content is. This section confirms the bots can get in.

RETRIEVABILITY (score /5)
[ ] 1. robots.txt does not block AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended)
[ ] 2. Key pages return HTTP 200 and are not blocked by meta robots noindex
[ ] 3. Content renders server-side or is available without JavaScript execution
[ ] 4. An llms.txt file exists and points AI models to your priority content
[ ] 5. XML sitemap is present, current, and submitted; canonical URLs are consistent

The most common failure here is an accidental block. Many sites disallow AI user agents in robots.txt without realizing it, which quietly removes them from ChatGPT and Perplexity. Confirm crawler access with the free robots.txt AI checker, and generate a discoverability file with the free llms.txt generator. JavaScript-only rendering is the second big trap: if your answer content only appears after a client-side fetch, many AI crawlers never see it.

Section 2: Answerability (5 checks)

Retrievable content still fails if an assistant cannot extract a clean answer from it. This is the section that most directly controls whether you get quoted.

ANSWERABILITY (score /5)
[ ] 6. Each key page has a 40-60 word quick-answer block in the first 100 words
[ ] 7. The page answers one clear query; the primary question is stated and answered directly
[ ] 8. Definitional clarity: key terms are defined in plain, self-contained sentences
[ ] 9. An FAQ section covers the real follow-up questions for the topic
[ ] 10. Answers are extractable without surrounding fluff (no burying the answer mid-paragraph)

Quick-answer blocks are the single highest-leverage AEO tactic, because assistants quote them close to verbatim. Write 40 to 60 words that name the subject and answer the title question directly, and place it in the first 100 words. Definitional clarity matters because AI answers are assembled from self-contained statements: a sentence that only makes sense with three paragraphs of context around it will not get lifted. Confirm your pages read as answerable with the free AI readiness checker.

Section 3: Structure (5 checks)

Structure is how you make an answer machine-readable. Schema, headings, and tables give assistants unambiguous signals about what your content means.

STRUCTURE (score /5)
[ ] 11. Relevant schema is present (FAQPage, Article, Organization, Product)
[ ] 12. Schema validates with no errors in a structured-data test
[ ] 13. Heading hierarchy is logical (one H1, H2s frame answerable sections)
[ ] 14. Data is presented in tables or lists where comparison or steps apply
[ ] 15. Entities (brand, product, people) are named consistently across the page

Schema is not optional for AEO. FAQPage and Article markup map directly to how assistants structure answers, and Organization schema anchors your brand as a recognized entity. Headings do double duty: a clean H2 structure gives both readers and models a table of contents, and each H2 should frame a question a user might ask. Tables win because comparison and step data is exactly what assistants pull into answers. Once your pages score well, compare AI visibility tracking tools to pick how you will measure the payoff.

Section 4: Authority (5 checks)

Assistants weigh trust before they cite. This section covers the signals that make an AI answer engine treat your page as evidence rather than noise.

AUTHORITY (score /5)
[ ] 16. Claims are backed by citations to authoritative external sources
[ ] 17. The page includes unique data, original research, or first-hand expertise
[ ] 18. Author or brand expertise is clear (bylines, credentials, About signals)
[ ] 19. Entity consistency across the web (NAP, brand name, descriptions match)
[ ] 20. Earned mentions and links from already-trusted third-party domains

Authority for AEO looks a lot like E-E-A-T for search, with an entity twist. AI assistants build a model of your brand from every place it appears, so inconsistent names, descriptions, or details across the web weaken your entity and your citation odds. Unique data is a citation magnet: assistants preferentially cite the original source of a statistic. Google’s own E-E-A-T guidance on experience, expertise, authoritativeness, and trust maps cleanly onto what answer engines reward.

Section 5: Measurement (5 checks)

You cannot improve what you do not track. This section confirms you have a feedback loop on whether AI assistants actually mention and cite you.

MEASUREMENT (score /5)
[ ] 21. You track brand mention rate across ChatGPT, Gemini, Perplexity, and Claude
[ ] 22. You track citation coverage (which domains AI answers link as evidence)
[ ] 23. You benchmark share of answer against named competitors
[ ] 24. You have a defined prompt set (category, comparison, solution) you retest
[ ] 25. You review AI visibility on a set cadence (monthly minimum, weekly in campaigns)

Measurement closes the loop. A mention (your brand named) and a citation (your domain linked) are different KPIs, and you want both. Define a prompt set of 50 to 200 queries per market, split across category, comparison, and solution intents, and retest on a schedule. Automate the capture with an AI mention tracker, and get a fast starting snapshot from the free AI visibility checker. AI answers are volatile, so a one-time score without ongoing tracking will mislead you.

Scoring Rubric

Add up your points across all five sections and read your band:

Score	Band	What it means
21-25	AEO-ready	AI assistants can crawl, extract, trust, and cite you. Maintain and monitor.
16-20	Solid, clear gaps	Fundamentals are in place. Close the failing checks to compete for citations.
11-15	Partially optimized	Real gaps in at least two sections. Prioritize retrievability and answerability.
6-10	At risk	AI assistants likely struggle to extract or trust your content. Fix crawler access first.
0-5	Invisible	Answer engines probably cannot use your content at all. Start at check 1.

Score your top revenue pages first and get them to 21+ before you work down the rest of the site. A handful of AEO-ready pages beats a whole site stuck at 12.

Turn the Audit Into a Program

A one-time audit is a snapshot. AEO is a moving target because assistants ingest the live web and models update, so treat this checklist as a recurring process: full 25-point audit quarterly, measurement check monthly, and a re-audit of any page after a substantive change. Wire the technical checks to the free tools above so anyone on your team can run them, and let the automated SEO audit handle the full sweep when you want the whole site scored at once.

Frequently Asked Questions

What is an AEO audit?

An AEO (Answer Engine Optimization) audit checks whether AI assistants like ChatGPT, Gemini, Perplexity, and Claude can crawl your content, extract clear answers from it, trust your authority, and cite you. This 25-point checklist scores five areas: retrievability, answerability, structure, authority, and measurement, giving you a 0-25 score and a prioritized fix list.

How is an AEO audit different from an SEO audit?

A traditional SEO audit checks whether Google can rank your pages: crawlability, keywords, links, and Core Web Vitals. An AEO audit checks whether AI answer engines can quote and cite you: crawler access for AI bots, quick-answer clarity, schema, entity consistency, and citation tracking. There is overlap on the technical basics, but AEO adds answerability and AI-specific measurement.

How often should I run an AEO audit?

Run a full 25-point audit quarterly, and a lighter measurement check monthly. AI answers shift constantly because assistants ingest the live web and models update, so a single snapshot goes stale fast. Re-audit any page after a major content change or PR push, and track your mention and citation rates continuously between full audits.

Can I do an AEO audit for free?

Yes. You can run every technical check in this list with free tools: the free robots.txt AI checker for crawler access, the free AI readiness checker for structure, the free llms.txt generator for discoverability, and the free AI visibility checker for mentions. For an automated full audit across all five areas, Rank Prompt’s SEO audit runs the checks and prioritizes fixes for you.

What score means my site is AEO-ready?

On the 25-point scale, 21 to 25 means AEO-ready, 16 to 20 means solid with clear gaps, 11 to 15 means partially optimized, 6 to 10 means at risk, and 0 to 5 means answer engines probably cannot use your content at all. Aim for 21+ on your priority pages first, then work down the rest of the site.

Which part of AEO matters most?

Retrievability comes first, because if AI crawlers cannot access your content, nothing else counts. After that, answerability tends to move the needle most: clear quick-answer blocks and definitional clarity are what assistants actually quote. Structure and authority compound those gains, and measurement tells you whether any of it is working.