How to Get Cited by ChatGPT, Perplexity & AI Search
A year ago, AI search handled a small fraction of English-language informational queries. By early 2026 that share had climbed sharply, and the curve isn’t slowing down. Here’s what that means for any brand spending real money on marketing: a growing share of your buyers are getting their answers from ChatGPT, Perplexity, Claude, Gemini, and Google’s AI Overviews, often without ever clicking through to a website. The question is no longer just “do we rank on Google.” It’s “when an AI engine answers a question in our category, does it cite us, or the competitor down the road?”
We work with founder-led eCommerce and high-value service brands across the UAE and Europe, and this is fast becoming one of the first things they ask us about. The discipline has a name now: Generative Engine Optimization, or GEO. It overlaps with traditional SEO, but the signals that earn you a citation in an AI answer are measurably different from the ones that earn you a blue link. Here’s what actually moves the needle, separated from the noise currently flooding LinkedIn.
Start with content, because that's what the research actually rewards
A lot of GEO advice treats AI citation as a formatting trick. It isn’t. The most credible evidence we have comes from the Princeton-led GEO study (Aggarwal et al., presented at KDD 2024), still treated as the benchmark in this space. Its finding was clear and a little inconvenient for anyone selling shortcuts: the methods that lifted visibility by 30 to 40% all involved adding genuine substance.
Three things worked consistently. Adding relevant, specific statistics improved visibility. Incorporating quotations from credible sources delivered similar gains. And citing authoritative references, real evidence rather than a tag declaring that something is a source, produced the same lift.
What didn’t work is just as instructive. Keyword stuffing performed worse than doing nothing. Writing in a more persuasive tone without adding any actual information showed minimal impact. The pattern is unmistakable: AI engines reward usefulness, not the appearance of it. If a tactic improves how your content reads to a machine without improving what it tells a human, it won’t hold up.
So before any technical work, audit your content for substance. Does each major claim carry a real number, a named source, or a genuine expert perspective? If your pages are full of confident assertions and empty of evidence, no amount of structural tinkering will make an AI engine trust them. This is exactly where most agency-produced “SEO content” falls down, and it’s the cheapest gap to close.
Make your content easy to extract
Once the substance is there, the structure decides whether an AI engine can actually lift and attribute it. Modern AI search runs on retrieval-augmented generation: the model pulls relevant passages from the live web, grounds its answer in them, and cites what it used. Your job is to make your passage the cleanest, most quotable option on the page.
Structure for machine extraction
A few structural choices carry outsized weight. Pages with proper H1, H2 and H3 nesting appear to be cited more often than unstructured content, because the heading hierarchy tells the engine what each section covers and where to find the answer to a specific query. Keep paragraphs tight, ideally under 120 words, one idea each. Self-contained chunks of 50 to 150 words tend to earn more citations than long, winding prose, because the model can extract them whole without losing the thread.
Use the formats AI engines prefer
Format matters more than people expect. Lists tend to get cited more often than the same information buried in paragraphs. Comparison tables outperform prose when the question is comparative. And well-placed pull quotes, especially ones that lead with a statistic, act as citation hooks: clearly demarcated, authoritative statements a retrieval system can grab and attribute.
One habit worth building in immediately: write your opening paragraph as a direct, complete answer to the question the page is about. AI engines frequently pull the first substantive passage, so leading with the answer rather than a throat-clearing introduction gives you a structural advantage from day one.
Build authority beyond your own website
Here’s the part most brands underinvest in, and where the real competitive gap opens up. The single strongest predictor of whether an AI engine cites you isn’t on your website at all. Across the current research, brand authority signals show the highest correlation with citation frequency, and engines build confidence by checking for agreement across multiple independent sources.
If your brand shows up consistently across industry publications, review sites, YouTube and community discussions, all describing you the same way, the model’s confidence in recommending you rises sharply. Sites cited across multiple AI platforms see higher appearance rates, and consistent entity information across the web has been associated with a meaningful lift in citation probability.
There’s also a clear preference for earned media. Third-party authoritative coverage tends to outweigh brand-owned content, particularly in ChatGPT, which rarely cites vendor blogs directly and leans on recognized publications and review sites instead. Social posts, by contrast, are almost entirely excluded from citations on most engines. So a serious GEO program isn’t just a content calendar. It’s digital PR, securing placements in the publications and directories AI engines already trust, earning real reviews on platforms like Clutch and Google, and making sure every one of those mentions describes you in the same words. For brands in the UAE and Europe, that also means getting your details consistent across regional directories, not just the global ones.
Get the technical foundations right, and skip the mythology
The GEO conversation has its share of overstated tactics sitting right next to a few that genuinely matter. Knowing the difference saves you a lot of wasted effort.
What’s overrated
The most overhyped item right now is the llms.txt file, a markdown file meant to tell AI crawlers which content matters most. It sounds tidy, but there’s currently no credible evidence it influences whether engines cite you, and Google’s John Mueller has publicly questioned tools that flag a missing one as a problem. Treat it as optional housekeeping.
Schema markup sits in a more nuanced place. Microsoft has confirmed Bing uses schema.org markup to interpret page content, and structured data like Article, FAQPage and Organization markup genuinely helps machines understand what your page is and who stands behind it. But research hasn’t found a direct link between schema coverage and citation rates on its own. The honest read: schema aids comprehension and entity disambiguation, worth doing, but it won’t manufacture citations without substance and authority behind it. Implement it, then move on.
What actually matters
What does demand attention is access. AI engines can’t cite a page they can’t read. Confirm the relevant crawlers, including GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot and Applebot, are allowed in your robots.txt, because one careless block can make you invisible to an entire platform. Make sure priority content renders without requiring JavaScript, since some AI crawlers don’t execute it, and keep load times under a couple of seconds. None of this earns a citation on its own, but any of it can quietly disqualify you.
It’s also worth knowing the engines behave differently. Google’s AI Overviews correlate strongly with traditional rankings, so classic SEO still feeds them. Perplexity weights freshness heavily and shows little source overlap between queries, giving newer, well-structured content a genuine shot against entrenched players. ChatGPT favors recognized brands above all else, which is why off-site authority pays off there more than anywhere. Because citation overlap across engines is low, monitoring one platform captures only a fraction of your real visibility.
The bottom line
Getting cited by AI engines comes down to a sequence that’s harder to fake than old-school SEO ever was. Earn the substance first, with real statistics, credible quotes and named sources. Structure it so a machine can extract a clean, self-contained answer. Build authority across the wider web so multiple independent sources agree on who you are. Clear the technical barriers that would otherwise hide you, and ignore the mythology that promises citations without the work. The brands that internalize this now will be the answer when their buyers ask. The ones that wait simply won’t exist in the conversation, and that gap is widening by the day.
At Social Schnell, AI-search visibility isn’t a bolt-on. We run it as part of one measured system alongside paid and SEO, so you can actually see what your marketing returns. The fastest way to find out where you stand is to ask the engines yourself: run your top category questions through ChatGPT, Perplexity and Google’s AI Overviews and see who they name. If it isn’t you, that’s the gap worth closing — and it’s exactly what we map in our free audit, the honest read on where your brand actually shows up before you spend another month’s budget trying to fix it.
Frequently asked questions
How long does it take to start getting cited by AI engines?
It depends on the lever. Structural fixes like rewriting opening paragraphs and cleaning up headings often show up in Perplexity within a week and in ChatGPT within a few weeks. Authority signals built through earned media and consistent mentions across the web typically take one to three months to register. GEO compounds over time rather than flipping like a switch, so treat it as an ongoing program, not a one-off project.
Is GEO replacing traditional SEO?
Not replacing it, standing alongside it. Google’s AI Overviews still lean heavily on traditional rankings, so strong SEO continues to feed AI visibility. What’s changed is that ranking on a results page is no longer the finish line. You also need to be the source an AI engine extracts and attributes, which requires a different set of signals layered on top of the SEO work you’re already doing.
Do I need schema markup to get cited?
It helps, but it isn’t the deciding factor. Schema like Article, FAQPage and Organization markup helps engines understand what your page is and who’s behind it, and Bing has confirmed it uses schema to interpret content. That said, research hasn’t found a direct link between schema coverage alone and citation rates. Implement it for clarity and entity disambiguation, then put your real effort into substance and authority.
Why does a competitor get cited when our content looks just as good?
Usually because of signals that live off your website. AI engines build confidence by checking whether multiple independent sources agree, so a competitor showing up consistently across publications, review sites and community discussions gets trusted more even with comparable on-page content. The gap is rarely the article itself; it’s the breadth and consistency of their presence across the wider web.
Does an llms.txt file actually help?
There’s no credible evidence that it does. The file is meant to tell AI crawlers which content matters most, but no major engine has confirmed using it to decide citations, and Google’s John Mueller has questioned tools that treat a missing one as a problem. It’s harmless to add, but it shouldn’t sit anywhere near the top of your priority list.
How do I know if AI engines are already citing my brand?
Test it directly. Run your most important category questions through ChatGPT, Perplexity, Claude, Gemini and Google’s AI Overviews, and note where you appear versus where competitors appear instead. Because overlap across engines is low, check several rather than one. Watching for AI referrers in your analytics, such as the perplexity.ai referrer, gives you a second data point as that traffic grows.
