AEO: how to get your brand cited by ChatGPT and Gemini

Ask ChatGPT where to buy the thing you sell. Go on. Type the question a real customer would type (“best accounting firm near Otopeni”, “who does server-side tracking for Shopify”, whatever your category is) and read the answer it hands back.

It names two or three companies, confidently. No ten blue links to scroll, no way to appeal. If your brand isn’t one of them, you just watched a sale happen on a shelf you didn’t know existed.

That shelf is the whole game now, and almost nobody is optimising for it. The discipline that gets you onto it has a name: Answer Engine Optimization, or AEO. It’s where I’d spend my next marketing hour before I touched anything else.

What AEO is

SEO gets you ranked. AEO gets you cited. Those are different jobs.

Classic search hands the user ten links and lets them choose. An answer engine (ChatGPT, Gemini, Perplexity, Google’s AI Overviews) reads the web on the user’s behalf, decides who’s credible, and returns one paragraph naming a few names. Position 3 stops mattering. You’re competing to be one of the two brands the model bothers to mention at all.

You’ll also hear GEO, Generative Engine Optimization. Same idea, slightly different emphasis. In practice I use the terms interchangeably, and so does most of the industry. Don’t let anyone sell you a course on the distinction.

The click is disappearing, and it isn’t coming back

Roughly 60% of Google searches now end without a single click (SparkToro and Datos, 2024: 59.7% in the EU, 58.5% in the US). The user got their answer on the results page (from a snippet, an AI Overview, a “people also ask” box) and never visited anyone’s site.

For twenty years the deal was simple: rank well, earn the click, get the visit. Now Google answers the question itself and keeps the traffic. If the answer doesn’t carry your name, you were never in the room. “We rank #1 for that” means less than it used to when the AI Overview above your result already answered the question.

Why the clicks that survive are worth more

When someone does click through from an AI answer, they arrive different. The model has already vetted you, framed you, half-recommended you. They’re checking out a name they were just handed by something they trust. In our experience that traffic converts 3-4× better than classic organic. Fewer visits, but far more of them ready to buy.

The click is rarer, and the one that survives is heavier. Being the brand the model names is the cheapest premium traffic of the decade, and the shelf is nearly empty. Most of your competitors haven’t noticed it exists.

What gets you cited

Nobody can see the model’s weights, so anyone promising a guaranteed spot in ChatGPT is guessing with confidence. What I can tell you is what consistently correlates with getting cited, from the accounts we measure:

• Entity clarity. The model needs to understand what you are, cleanly. That means schema markup, a consistent set of brand facts, and an entity page it can read and trust: who you are, what you do, where, since when, who runs it.
• The same facts everywhere. If your founding year, your location or your service names differ across your site, your Google profile, LinkedIn and directories, the model hedges. Contradiction reads as low confidence. Boring consistency wins.
• A real founder page. Models weight named, accountable humans. A page with a real person and a real track record does more for your credibility with a model than another anonymous blog post ever will.
• Presence where the models learn. They don’t read your site in a vacuum. They read the sources they already trust and infer you from those. Getting mentioned in the places they cite is half the work.
• Answering real questions in plain language. The content that gets quoted answers the question a customer actually asked, in their words, near the top, without ten paragraphs of throat-clearing first.

You can’t optimise what you can’t see

Most “AEO services” skip measurement because it’s the hard part. If you’re not reading what the engines say about your category, you’re decorating.

We built our own AI visibility engine for exactly this. It asks the five main engines the questions your customers ask, then measures three things: your mention rate (how often you come up at all), your share of voice against named competitors, and an AEO score from 0 to 100. Each one is backed by the real answers and citations, so you can read the receipts and argue with the verdict. Our founder built it. If you want your numbers, start with an AI visibility audit. It’s how you catch a competitor quietly becoming “the answer” in your category while your rankings still look fine.

Why now, specifically

This looks like SEO in 2004. Back then a bit of structure and a few good links could put a small business above incumbents who assumed search didn’t matter. That window closed years ago. The AEO window is open right now, and it’s wider, because the models are hungry for clear, credible sources and most categories have almost none.

In two years, “how do we show up in ChatGPT” will be a line on every marketing brief. By then the brands that started early will already be the answer, and the model will have spent two years learning to say their name. First-mover advantage compounds here, because the model’s confidence in you compounds.

Where to start

We measure where you stand in AI answers today, with evidence, then go earn you the citation. You can see how the work is structured on the AEO and AI visibility page, or skip to the number that matters and book an AI visibility audit. We’ll show you exactly what ChatGPT and Gemini say about your category right now, competitors included.

One caveat I’ll repeat until it sticks: none of this replaces getting your own house in order. Clean tracking, clear facts, a real entity behind the brand. Models reward the same things good customers do, which is the part I like about this.