To get cited by ChatGPT: make sure AI crawlers can access your site (robots.txt), hand them a clean summary of what you do (llms.txt), answer questions immediately instead of making people scroll, structure your content with proper headings and schema, and show up on platforms like Reddit and LinkedIn, because that's where AI is already eavesdropping.
Okay, story time.
A while back, a now-client of ours asked ChatGPT a question that their product literally, directly, answers.
Their name didn't come up. Their competitor's did. Twice.
Rude, honestly. But also, extremely useful, because it made them realize something: ranking #1 on Google and getting cited by ChatGPT are two completely different games. And most brands are only showing up to play one of them.
If you're reading this, you've probably clocked the same shift. People aren't Googling and scrolling through ten blue links like it's 2015 anymore. They're asking ChatGPT (or Perplexity, or Gemini) a question and just... taking the answer. No click-through. No page visit. Just vibes and an AI-generated paragraph.
So the real question isn't "how do I rank." It's "how do I get picked." Slightly more existential, we know. Stay with us.
Google finds you. AI has to actually trust you.
Google's whole job is matching. It reads your page, matches it to a search query, ranks it, done. Very transactional. Very "we're not friends, we're just doing business."
AI engines are needier. They have to read your content, decide if it actually answers the question, and then decide whether they trust it enough to lift a piece of it and hand it to someone, sometimes with your brand's name attached, sometimes not (rude, again).
Most content on the internet is written to be found. Almost none of it is written to be quoted.
That gap is basically the entire job description of AEO (Answer Engine Optimization), which, yes, is a real acronym now, and yes, we also can't believe marketing needed another one.
The unglamorous stuff first: robots.txt and llms.txt
Before we even get to writing anything, there's a boring little gatekeeping layer almost nobody checks.
robots.txt is a tiny file sitting at the root of your website that tells any crawler (Google's, Bing's, ChatGPT's) what it's allowed to look at. If it's quietly blocking AI crawlers, sometimes by accident, sometimes as a fossil from an old developer's decision three redesigns ago, it genuinely does not matter how good your content is. The AI never even sees it.
Go check yours. Type yourdomain.com/robots.txt into your browser right now. You might be mildly betrayed by what you find.
llms.txt is the newer cousin of this idea, a file that hands AI models a clean, organized summary of what your site actually is and where your best content lives. Think of it as a cheat sheet you're leaving out specifically for the AI, instead of making it guess.
Almost nobody has an llms.txt yet. Which, honestly? Great news for you. That's a gap, not a requirement, which means being early here actually means something.
Neither file guarantees a citation. But skip them, and you're not even in the running.
Answer the question. Immediately. We mean it.
This is the one shift that matters more than almost everything else on this list.
Think about how most articles are written: a long, scenic intro, some context, a little story (we're doing it right now, we know), and then, eventually, the actual answer, three paragraphs deep.
That worked when a human was settling in to read the whole thing. It does not work when an AI model is skimming your page in half a second trying to find something quotable.
- Buried answer no pickup
- Answer sitting right at the top in plain language big pickup
Say the actual answer in the first line or two, phrased so it could stand alone if someone yanked it out of context completely. Then go ahead and tell your story, add the nuance, do the scenic route for the humans who stuck around.
It'll feel almost too blunt when you first write it that way. That's usually the sign you did it right.
Structure your site like you're explaining it to a machine, not a mood board
Here's the plot twist: a page can look gorgeous to a human and still be an absolute mess to an AI crawler trying to parse it.
Heading hierarchy. Your H1 should say plainly what the page is about, not a vibe, not a clever tagline, not a pun (we love a pun, but not here). Your H2s should break the page into clear sections, each one answering a sub-question. Basically: your headings are a table of contents you're handing straight to the machine. If someone could skim just the headings and get the gist of the whole article, you nailed it.
Schema markup. This is structured data in your site's code that flat-out tells search engines and AI models what things are: this is a product, this is a price, this is an FAQ, this is a how-to step. Without it, the crawler is guessing from context, like reading a group chat with no punctuation. With it, you've just handed over the answer key.
FAQ schema specifically is worth prioritizing, it's already in the question-and-answer shape AI engines love to pull directly from.
None of this is glamorous. Nobody's screenshotting your schema markup for their moodboard. But it's often the exact difference between a crawler understanding your page in one pass, and bouncing off it with nothing to hold onto.
Go where the AI is already eavesdropping
Here's the part most brands completely miss: your website isn't the only place AI is learning about you. Honestly, it might not even be the main one.
Certain platforms get pulled into AI-generated answers constantly, real threads, real conversations, real people asking and answering the exact questions your audience is asking. Reddit. LinkedIn. Niche forums for your specific industry that you didn't even know existed until you needed them to.
Why does this happen? Because AI models weight unscripted, real human discussion pretty heavily. It reads as more trustworthy than polished brand copy. Because, let's be honest, it usually is.
So part of your AEO strategy legally cannot just live on your own domain. It looks more like:
- Actually being useful in relevant subreddits and forums, answering like a person who knows things, not dropping a link and running
- LinkedIn content from real humans, not just the company page. Personal accounts consistently outperform brand pages here
- Getting mentioned or linked to in the places your actual audience already hangs out, unprompted, because you said something worth repeating
If your brand only exists on your own website, you're giving AI one tiny window into who you are. The brands getting cited consistently are the ones showing up everywhere the AI is already listening.
The actual takeaway
Getting cited by ChatGPT isn't about gaming an algorithm. It's about making it stupidly easy for an AI model to find you, understand you, trust you, and quote you, technically, structurally, and across all the places real conversations about your industry are already happening.
Most brands haven't even started. Which, selfishly, is kind of great news for the ones who do.
If you want a straight answer on whether AI engines are actually picking up your brand right now spoiler: most brands are surprised, and not in a fun way get on a call with us. We'll show you exactly where you stand and what we'd fix first.