AEO Citation Checklist: What to Ship Before You Get Cited
First citations from ChatGPT, Perplexity, Gemini, and Claude tend to land 30 to 90 days after a site reaches a baseline of AEO readiness. The 20 items below are what I treat as that baseline, in priority order, with the rationale per item. The conceptual primer is what is AEO; the implementation walkthrough is how to get cited by ChatGPT.
I am Ignacio (IGNAX), Spain-born and Paraguay-based, a solo full-stack developer and SEO/AEO specialist. I am running this exact checklist on ignax.dev itself; the live experiment is the growth engine case study.
What is the short checklist?
Grouped by category, in ship order:
Content (P0, ship first):
- Quick Answer block (40–60 words) on every key page.
- All H2s phrased as questions on long-form pages.
- FAQ block with 5–8 Q/As on every page.
Schema (P0):
4. FAQPage JSON-LD validated in Rich Results Test.
5. Article JSON-LD with author, datePublished, dateModified, inLanguage.
6. Service JSON-LD with provider, areaServed, offers on service pages.
7. Person JSON-LD with full sameAs array on /about.
8. Organization JSON-LD with sameAs and logo.
9. BreadcrumbList on every nested page.
Discoverability (P0):
10. llms.txt at the domain root, curated.
11. llms-full.txt at the domain root, auto-generated.
12. sitemap.xml walking the content registry, with hreflang annotations.
13. robots.txt allowing GPTBot, ClaudeBot, PerplexityBot, Google-Extended.
14. IndexNow integration pinging on every publish.
Identity (P1):
15. Person sameAs includes LinkedIn, GitHub, X, and any reputable identity provider.
16. Organization sameAs includes company LinkedIn, X, Crunchbase if applicable.
Multilingual (P1, conditional): 17. Hreflang siblings paired across all locales. 18. Locale-specific Quick Answers (not machine translations).
Performance (P1): 19. Lighthouse green (LCP < 2.5s, INP < 200ms, CLS < 0.1) on every key template. 20. GSC and Bing Webmaster Domain properties verified, sitemap submitted.
That is the entire checklist. The rationale per item follows.
Why is the Quick Answer block #1?
Quick Answers are the single highest-leverage AEO change. AI engines lift them verbatim into responses far more often than they lift deeper paragraphs. The rules:
- 40 to 60 words (~150 to 450 characters).
- In the first 200 pixels of the rendered page.
- No anaphora: lead with the topic, not "this article".
- Direct factual answer to the page's target query.
- No promotional language.
For more on writing Quick Answers see how to get cited by ChatGPT. The schema-enforced 150–450 character range I ship is encoded in the front-matter Zod validation so drift fails the build.
Why H2s as questions?
AI engines parse Q-as-heading structure efficiently. A user query "what is AEO" matches more cleanly to an H2 reading "What is AEO?" than to one reading "AEO Overview". The match increases the probability the section gets quoted.
The discipline:
- Every
##heading is phrased as a question on long-form content. - Headings on service pages mirror common buyer questions ("Who is this for?", "What does it cost?", "How does it work?").
- Avoid generic headings ("Introduction", "Benefits", "Conclusion") on AEO-targeted pages.
This article uses the pattern throughout. Every H2 above is a question.
Why FAQPage schema?
FAQPage is the highest-leverage schema type for AEO citation. AI engines lift FAQ answers into responses verbatim when the question matches the user query.
Implementation rules:
- 5 to 8 questions per page (more dilutes the signal).
- Each
textanswer ≤80 words (~640 characters). - Questions phrased the way a real user would type them.
- Validated in Google Rich Results Test.
The schema lives in JSON-LD; the visible UI matches the schema content. For the implementation pattern see llms.txt explained which uses the same registry-driven approach.
Why llms.txt and llms-full.txt?
llms.txt is a curated index of your high-value URLs with one-line descriptions. llms-full.txt is the same set of URLs with full markdown content concatenated. Both files live at the root and act as a table of contents for AI crawlers.
The cost of shipping is near zero when auto-generated from the content registry. The expected value is positive even if only one engine respects the spec. For the full walkthrough see llms.txt explained.
Why IndexNow?
IndexNow notifies Bing, Yandex, and aggregators of new or updated URLs within seconds of publish. Bing's index feeds ChatGPT browsing; faster Bing crawl means faster ChatGPT awareness.
The integration is one API call per publish. See IndexNow setup tutorial for the full walkthrough.
Why Person sameAs?
The sameAs array verifies you are a real entity, not a name collision. AI engines use sameAs to disambiguate when asked about a specific person or company.
Include in sameAs:
- LinkedIn profile.
- GitHub profile.
- X (Twitter) profile.
- Crunchbase or AngelList profile (if applicable).
- ORCID (if you are an academic).
- Conference speaker pages.
- Podcast guest pages.
- Personal blog.
The more entries in the array. The more confidently the AI engine can attribute. For ignax.dev the array currently has six entries and grows over time as I appear on podcasts, guest posts, and directories.
Why Lighthouse green?
Core Web Vitals affect crawl frequency. Faster sites get crawled more often, which means new content reaches AI engines faster. Slow sites get crawled less and trust the data less.
Targets:
- LCP under 2.5s (synthetic and 75th percentile real-user).
- INP under 200ms.
- CLS under 0.1.
For the implementation playbook see Core Web Vitals on Cloudflare Pages.
What is the weekly maintenance habit?
After shipping the 20-item checklist. The AEO surface needs maintenance:
- Monday: GSC pull. Check rising queries, falling CTR queries, indexing errors.
- Tuesday: Bing Webmaster check. Mirror the GSC review.
- Wednesday: Schema spot check. Rich Results Test on any recently changed templates.
- Thursday: Publish or rewrite. One to three content updates per week minimum.
- Friday: AI citation probe. Run 5 to 10 target prompts across ChatGPT, Perplexity, Gemini, Claude. Log results.
Sixty to 90 minutes per week. Sites that run it consistently see citation growth compound; sites that skip the loop watch citations stagnate. That habit is what separates "AEO ready" from "AEO citing".
For the full weekly loop methodology see the growth engine case study.
What does the 90-day timeline look like?
Realistic expectations from a clean checklist ship:
| Day | Milestone |
|---|---|
| 0 | All 20 items shipped, validated, deployed |
| 1–14 | GSC + Bing index every URL via IndexNow pings |
| 14–30 | Rankings appear on long-tail keywords |
| 30–60 | First AI engine citations on niche queries |
| 60–90 | Citations on competitive buyer-intent queries |
| 90+ | Compounding, earlier citations lift later citations |
The 30 to 60 day window for first niche citations is real. I have seen it on multiple bilingual SvelteKit sites. The 90-day window for competitive citations is also real but variable, depends on how dense the topical authority graph already is.
How do I validate the checklist?
Per-page scorecard:
| URL | QA block | FAQ schema | H2-Q | Person sameAs | llms.txt entry | Score |
|---|---|---|---|---|---|---|
| / | ✓ | ✓ | ✓ | ✓ | ✓ | 5/5 |
| /services/saas-mvp | ✓ | ✓ | ✓ | ✓ | ✓ | 5/5 |
| /about | ✓ | ✓ | ✓ | ✓ | ✓ | 5/5 |
| /articles/what-is-aeo | ✓ | ✓ | ✓ | ✓ | ✓ | 5/5 |
Pages scoring under 4/5 get prioritized in the next sprint. For the audit-grade version of this scorecard see technical SEO audit deliverables.
What external references should I read?
- Google Search Central structured data documentation: canonical JSON-LD reference.
- Schema.org full type list: every schema type, in one place.
- llmstxt.org; the llms.txt proposed standard.
- IndexNow protocol: fast-indexing API spec.
- OpenAI GPTBot documentation: how ChatGPT's crawler behaves.
Ship the 20-item checklist. Run the weekly loop. Expect first citations in 30 to 90 days. The work is mechanical once the pattern is in place, see the SEO + AEO setup service for engagement details.
Ready to ship the checklist? Email hello@ignax.dev with your URL. I run the crawl before our call.
Frequently asked questions
Which item on the checklist has the highest leverage?
The Quick Answer block. A 40 to 60 word stand-alone factual paragraph in the first 200 pixels of every key page produces the largest citation rate improvement per hour of work. AI engines lift Quick Answers verbatim into their responses. Everything else (schema, llms.txt, IndexNow) compounds the effect but the Quick Answer is the single highest-ROI change you can ship in one afternoon.
Can I skip any items?
Skip Organization schema if you are a solo entity (use Person only). Skip hreflang if you are single-language. Skip llms-full.txt if your llms.txt is comprehensive enough. Beyond those three exceptions, every item on the checklist earns its place. Skipping items because they look optional is the most common AEO mistake. The checklist is short on purpose; trim with caution.
How long does shipping the full checklist take?
Five to ten days on an existing site, depending on size and content volume. A fresh site with the AEO playbook baked into the build ships the checklist on day one without retrofitting cost. Most of the time goes into rewriting existing content to have Quick Answers and FAQ blocks, not into the technical schema work. For an existing 50-page site expect roughly 30 to 60 hours of work total.
Is there a minimum publish cadence?
One published or updated piece of content per week, minimum, to stay visible to AI engine recrawl cycles. Two to three per week is the sweet spot, enough signal that engines crawl regularly without burning out your authoring time. Sites publishing once a month tend to lose citation share to sites publishing weekly because the engines crawl less and trust the data less.
How do I know when I'm ready to expect citations?
When all 20 items are shipped and validated, and the weekly publish cadence has been running for at least 4 weeks. Earlier than that you cannot distinguish between 'AEO not working' and 'AEO has not had time to work'. Set a 90-day milestone for first citations on niche queries and a 180-day milestone for competitive queries. Track via the manual citation probe documented below.
What is the manual citation probe?
A weekly habit: pick 10 to 20 target prompts the way a buyer would type them, run them across ChatGPT (browsing on), Perplexity, Gemini, and Claude, log whether your domain was cited. After 8 to 12 weeks you have enough data to see the trend. The probe is unavoidably manual in 2026; no public API exposes AI engine citations. Treat it as the GSC equivalent for AEO and build the habit into the weekly maintenance loop.