GEO/AEO checklist: how to get cited in AI answers
Seven concrete steps that increase the chance of AI systems like ChatGPT, Perplexity or Google AI Overview citing you as a source. No theory — just what to do and why.
I have been doing classical SEO for years. Adding keywords, building backlinks, tuning Core Web Vitals. It works. But in the last few months a new channel has appeared that does not work quite the same way: generative AI.
ChatGPT, Perplexity, Google AI Overview — all of these systems answer queries by citing sources. Either they cite you, or they do not. It is not just about your position in search results. It is about whether your content can be read, understood and used as an answer by AI.
Here is a checklist of seven things that influence that. For each one I explain what and why.
1. llms.txt and allowing AI bots in robots.txt
The llms.txt file in the root of your site is a convention (inspired by robots.txt) that tells AI crawlers what your site does, who is behind it and what content it offers. It is not a standard — but Anthropic, OpenAI and Perplexity all follow it.
Also check your robots.txt. If you are blocking GPTBot, Claude-Web, PerplexityBot or GoogleOther, AI cannot see your content at all.
2. Structured data JSON-LD
Add Schema.org schemas to your pages — specifically Organization, Article and FAQPage. AI systems (especially Google AI Overview) actively use this structured data to identify the author, the context of the article and ready-made Q&A pairs.
Without JSON-LD, AI has to infer structure from raw HTML. With correctly set up schemas it gets the information directly.
3. Content structured as answers to queries
AI looks for answers in content, not texts. Every H2 or H3 should be phrased as a question or a direct statement that answers one. Add a FAQ section with concrete Q&A pairs — Perplexity and Google AI Overview cite it preferentially.
Instead of "About our service" write "What our KSeF integration can do". Instead of a general introduction, start with the answer.
4. Concrete numbers and facts
AI cites numbers, not adjectives. "Performance improved significantly" leaves no trace. "We cut the dataloader from 20 hours to 30 minutes" is a citable fact.
Put concrete metrics, timeframes, record counts and percentage changes into your content. Numbers act as anchors — AI picks them up exactly.
5. Speed and technical SEO
A slow site gets skipped by AI crawlers or indexed with a delay. Core Web Vitals (LCP under 2.5 s, INP under 200 ms, CLS under 0.1) are the minimum. If your page uses JavaScript rendering, add prerender or SSR — AI bots do not execute JavaScript.
Server-side rendering or a static build are clearly better than a client-side-hydrated SPA from the perspective of AI readability.
6. Multilingual hreflang for international reach
If you are targeting multiple markets, hreflang attributes tell AI which language version is relevant for a given query. Without hreflang, Perplexity or Google AI may serve the Czech version to a German-speaking user — and skip the citation.
A trilingual site (CS/EN/DE) with correct hreflang covers queries from different language markets without duplicate indexing.
7. Measure via Google Search Console
GSC does not show AI citations directly, but the signals are there. Watch branded queries — if new visitors start appearing who search for your brand by name, it is likely they saw you somewhere (in an AI answer, in an overview). Also watch CTR on pages with FAQ schemas.
Direct measurement of AI impressions is still coming. GSC is the closest available tool for now.
What we did on this site
On solutionbox.cz we have all seven points in place. llms.txt and allowed AI bots, JSON-LD schemas Organization, Article and FAQPage on every relevant page, content structured as questions and answers, concrete numbers from projects (15,141 documents, 20 hours to 30 minutes, thirty million invoices per year), prerender with 0 ms TTFB, trilingual CS/EN/DE with hreflang, and GSC connected to search analytics.
I am not mentioning this to sound impressive — but because if you are solving these things on your site, we know how to do it.
FAQ
What is GEO and AEO?
GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) are approaches to optimizing a website so that generative AI systems like ChatGPT, Perplexity or Google AI Overview cite it as a source. They extend classical SEO with technical signals that make content readable for AI.
How do I find out if AI is citing me?
There is no direct tool for tracking AI citations yet. An indirect signal is Google Search Console — watch branded queries and backlinks. You can test manually: enter a query into Perplexity or ChatGPT and check whether your site appears among the sources.
Do I need llms.txt for AI to cite me?
llms.txt is not a guaranteed path to citation, but it is a readable signal for AI crawlers that tells them how your site works and what it offers. Combined with other steps (structured data, concrete facts, speed) it raises the overall AI-readability of your site.