The 25-Point AEO Checklist That Grades Your AI-Search Readiness

By the Lenoretech SEO Strategy Team · Reviewed by a senior SEO strategist · Last updated: June 2026

An AEO checklist audits whether AI answer engines can extract, trust and cite your content. Score 4 points for each of the 25 items below to get a 0-100 AEO readiness grade: a score of 0-49 means engines rarely cite you, 50-74 means occasional citations on lower-competition queries, and 75 or above means you are a default source AI reaches for. Most sites we audit score between 28 and 44 on the first pass, even when their Google rankings are healthy.

The reason is simple. Classic SEO optimises a page to win a click. AEO optimises a passage to be quoted inside an answer the user never clicks away from. Different signals, different formatting, different success metric. The checklist below is grouped into the four signal families that decide citation likelihood: answer formatting, entity clarity, machine-readable infrastructure, and citability or trust signals. Score honestly, total it up, and you will know precisely where to spend your next sprint.

How the scoring works

Each item is worth 4 points: give yourself the full 4 if it is fully done across your priority pages, 2 if it is partly done, and 0 if it is missing. Add the 25 scores for a total out of 100. We use these four thresholds from real client tracking across ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews:

• 0-49 (Invisible): AI engines can read your page but rarely choose it. Expect citations only on near-zero-competition queries.
• 50-74 (Emerging): You get cited on long-tail and branded queries. Inconsistent on commercial terms where competitors are stronger.
• 75-89 (Citable): You are a regular source in your niche. AI Overviews quote you and Perplexity links you on most relevant prompts.
• 90-100 (Default source): Engines reach for you first. You are quoted, named as an entity, and used to fact-check competitors.

Section A - Answer formatting (items 1-7)

AI engines extract passages, not pages. If your answer is buried three scrolls down inside a 400-word preamble, the model often skips it for a competitor who answered in the first sentence. This section is the fastest to fix and usually the biggest score jump.

• 1. Direct answer in the first 1-3 sentences. Every key page opens by answering the query before any setup. This single habit moves more of our clients into AI Overviews than any other change.
• 2. Question-shaped H2s and H3s. Headings phrased the way people actually ask ("How much does X cost?") match retrieval queries far better than clever marketing headlines.
• 3. Self-contained passages. Each section makes sense if quoted alone, with no "as mentioned above" dependencies that break when extracted.
• 4. Specific numbers, dates and ranges. "Most see results in 4-8 weeks" gets quoted; "results vary" never does. Concrete data is the most-cited content type we track.
• 5. Lists and tables for comparable facts. Steps, pros and cons, pricing and specs in real <ul> or <table> markup are extracted cleanly into answers.
• 6. A genuine FAQ block with schema. Five or more real questions with tight 40-60 word answers feed both FAQ rich results and AI answer panels.
• 7. Scannable formatting density. Short paragraphs, bolded key terms and clear sub-sections. Walls of text get summarised away, not quoted.

Section B - Entity clarity (items 8-13)

AI engines model the world as entities and relationships, not keywords. If a model cannot confidently identify who you are, what you do and how you connect to known topics, it will not put your name in an answer it is willing to stand behind. This is the section most agencies skip, and it is the difference between being read and being named.

• 8. A clear, consistent entity name. Your brand, people and products are referenced the same way everywhere, so the model resolves them to one entity instead of three fuzzy ones.
• 9. Organization and Person schema deployed. Structured sameAs, founder, and knowsAbout properties tell engines what you are an authority on. See our AEO services for the full schema stack we deploy, and our 15 schema markup examples for copy-paste templates.
• 10. An entity home page (About). One canonical page that defines the entity, with credentials, history and scope, that everything else links to.
• 11. Wikidata and Knowledge Graph presence. For established brands, a Wikidata entry and consistent name, address and phone across the web build the entity an engine can trust.
• 12. Topical clusters, not orphan posts. A pillar page plus tightly linked supporting articles signals depth. Read our take on building topical authority without backlinks.
• 13. Explicit relationships in copy. State plainly how your product relates to its category, competitors and use cases. Engines extract relationships that are written down, not implied.

Section C - Machine-readable infrastructure (items 14-19)

You can write the perfect answer and still lose if a crawler cannot fetch it cleanly or a model is blocked from your content. This section is the plumbing.

• 14. Server-rendered content. Key answers exist in the raw HTML, not injected by JavaScript that some AI crawlers never execute. This is a silent killer for many React and Vue sites.
• 15. AI crawlers allowed in robots.txt. GPTBot, PerplexityBot, ClaudeBot, Google-Extended and OAI-SearchBot can reach the pages you want cited. Blocking them by accident is common.
• 16. An llms.txt file. A root-level /llms.txt that points models to your most important, citation-worthy pages in clean markdown. Early, but cheap insurance and increasingly read.
• 17. Valid, error-free schema. Article, FAQ, Product, HowTo and Breadcrumb markup that passes validation. Broken schema gets ignored.
• 18. Fast, stable Core Web Vitals. Slow or unstable pages get crawled less and re-fetched less, which ages out your content from retrieval indexes.
• 19. Clean canonical and sitemap signals. One canonical version, a current XML sitemap, and no duplicate-content noise confusing which page is the source of truth.

Section D - Citability and trust signals (items 20-25)

The last six items are what tip a model from "could cite" to "will cite". Engines weight sources that demonstrate first-hand experience and verifiable authority, because their own quality systems are tuned to avoid quoting thin, anonymous or recycled content. In our audits, two sites with identical formatting scores often diverge by 20-plus citation appearances purely on the strength of this section. Trust is the multiplier on everything above it.

• 20. First-hand experience and original data. Proprietary numbers, case results, screenshots, tests or surveys that exist nowhere else. Original data is the single strongest citation magnet we track, because a model cannot synthesise it from competitors and must name you as the source.
• 21. Author bylines with real credentials. Every key page carries a named author, a short bio, and links to their profiles and expertise. Anonymous "admin" posts are routinely down-weighted; a credentialed byline tied to Person schema tells the engine a human with standing stands behind the claim.
• 22. Citations to primary sources. Link out to studies, official docs and original research with dates. Counterintuitively, citing strong sources raises your own citability, because models trust pages that sit inside a verified evidence chain rather than asserting facts in a vacuum.
• 23. Third-party reviews and reputation signals. Verified reviews, ratings, testimonials and mentions on sites you do not control. Engines cross-check whether the wider web agrees you are credible before naming you, so off-site reputation directly feeds on-page citability.
• 24. Freshness and visible last-updated dates. A real "last updated" date and genuinely refreshed content. For fast-moving topics, engines strongly prefer recent sources; a page last touched in 2023 loses to a 2026 equivalent on almost every time-sensitive query.
• 25. Consistent off-site mentions. Guest posts, podcasts, directory listings and unlinked brand mentions that repeat your entity name and expertise. The more the web corroborates who you are, the more confidently an engine will quote you by name.

Total your score and pick your next move

Add all 25 item scores for your grade out of 100. If you landed in the 0-49 band, start with Section A: fixing direct answers and adding genuine FAQ blocks is the cheapest path from invisible to occasionally cited, and it usually adds 15-20 points in a single sprint. If you are 50-74, your formatting is probably fine and your gap is entity clarity (Section B) or trust (Section D); that is where mid-scoring sites plateau. Sites above 75 should protect their lead with freshness, original data and llms.txt coverage on every priority page.

One honest caveat from running these audits: AEO and traditional SEO are not opposites. The same server-rendered, well-structured, authoritative page that gets cited by ChatGPT also tends to rank better on Google, and the same E-E-A-T signals power both. If you want the underlying playbook for shaping passages that get quoted across engines, read our guide to optimising content for ChatGPT, Perplexity and Google AI Overviews, and pair it with broader generative engine optimization work if AI traffic is now a real channel for you. Re-score quarterly, because the bands move as engines tighten their quality bars.