AEO Guide: How to Get Cited by AI in 6 Steps

Answer Engine Optimisation (AEO) is the practice of structuring your content so AI tools like ChatGPT, Gemini, Perplexity, and Claude cite your business when they answer questions. There are six core steps: make your site readable by AI, write answer-first content, keep your brand entity consistent, build credible off-site mentions, add structured data, then refresh and measure. Research shows adding statistics and citations can lift AI visibility by up to 40%.

Your buyers are asking AI assistants questions before they ever open a browser tab. Google’s AI Overviews now reach roughly 1.5 billion monthly users, and at its May 2026 I/O event Google confirmed that AI Mode passed one billion monthly users within a year of launch, with queries more than doubling every quarter. The question isn’t whether your customers use AI search. It’s whether your business shows up in the answer.

Here’s what actually moves the needle.

Key takeaways

• Allow AI crawlers, keep important content in plain HTML, and consider an llms.txt file, or none of the rest matters.
• Write answer-first content with clear question headings and a crisp reply directly underneath.
• Back claims with statistics and citations. The Princeton GEO study found this lifts AI visibility by up to 40%.
• Keep your brand entity identical everywhere, and earn mentions on the sites AI engines actually cite.
• Refresh content regularly and benchmark monthly. Freshness is a measurable ranking factor.

Step 1: Make sure AI can read you

Before anything clever, check the basics. AI engines can only cite content they can fetch and parse. Allow the major crawlers in your robots.txt: GPTBot and OAI-SearchBot (OpenAI), PerplexityBot, ClaudeBot (Anthropic), and Google-Extended. Then make sure your key content lives in crawlable HTML, not locked behind JavaScript, because some LLMs cannot render JavaScript. Speed helps too. One analysis found fast-loading pages were cited roughly three times more often than slow ones. If your most important answers are buried in a slow, script-heavy page, you’ve lost before you’ve started.

You can also add an llms.txt file, a simple text file in your site root that points AI tools to your most important content in clean markdown. Be realistic about it, though. Google has said llms.txt isn’t required for AI search visibility, and there’s no confirmation the major engines use it for citations yet. Treat it as a low-effort, forward-looking step rather than a guaranteed win. It takes minutes to add and may help as adoption grows, but it won’t move the needle on its own.

Step 2: Write answer-first content

AI engines pull from content that answers a question directly. Structure your pages with clear question-based headings, then answer immediately underneath with no preamble. Think “What does [your company] do?” and answer it in one clean sentence a model can lift whole. Phrase those headings the way people actually ask, in full conversational questions rather than short keywords, because that’s how queries arrive in AI tools.

This is also where most businesses underperform, because they don’t include anything worth citing. The foundational research here is the Princeton GEO study, presented at ACM KDD 2024, which tested nine content tactics across 10,000 queries. Five of them boosted AI citation rates by 30 to 41%: citing sources, adding quotations, adding statistics, improving fluency, and using an authoritative voice. Keyword stuffing and promotional language did nothing or actively hurt. So write with specifics. “Companies using this approach saw a 67% faster rollout” beats “companies rolled out faster” every time, and a factual tone tends to land better than marketing copy.

Step 3: Make your entity consistent everywhere

AI systems build a picture of your brand from dozens of sources. Your name, description, services, and location need to match across your website, Google Business Profile, LinkedIn, G2, and any directory you appear in. Inconsistency creates doubt, and AI engines resolve doubt by citing someone clearer.

Step 4: Build credibility on the sites AI actually reads

Off-site mentions tell AI engines that real people endorse you, but it helps to know where these tools look. ChatGPT leans heavily on Wikipedia, Reddit, and Forbes, while Perplexity and Google AI Overviews pull disproportionately from Reddit, YouTube, and Quora. Reviews on G2 and Google, genuine contributions on Reddit and industry forums, and coverage in relevant publications all feed the picture. Brand mentions matter more than you’d expect. One analysis found brand mentions correlate with AI citation far more strongly than backlinks. Ask your best clients for reviews, and contribute real answers where your buyers are already asking questions.

Step 5: Use structured data to remove ambiguity

Schema markup (Organisation, Service, FAQ, Review) tells crawlers what your content means, not just what it says. It won’t guarantee a citation, and Google has said it isn’t strictly required for AI visibility, but it removes ambiguity and helps extraction. If your site uses no schema at all, this is a quick win worth prioritising.

Step 6: Refresh, measure, and iterate

AEO isn’t set-and-forget. AI engines favour recency. Cited URLs tend to be about 25.7% fresher than those ranking in traditional search, so revisiting key pages quarterly keeps you in the running. For measurement, run a baseline with a free checker like HubSpot’s AEO Grader, an entry-level tracker like Otterly.AI, or an enterprise platform like Profound, then track monthly whether your brand appears when you ask AI tools the questions your buyers ask. When you show up, note what content got cited. When you don’t, that’s your next content gap.

There’s a real first-mover advantage here worth acting on. HubSpot’s research found that over 92% of marketers plan to or already optimise for traditional and AI-powered search, yet far fewer have actually built an AEO strategy. The gap between intent and action is your window. One more thing worth knowing: the strongest way to earn citations is to publish something only you have, original data, a documented client result, a comparison you actually ran, since AI engines have little reason to cite yet another summary of advice already on the web. If you’d like a hand auditing where your brand stands across the AI engines, get in touch with the Cybersolve team.

Frequently asked questions