
Just like traditional SEO, 90% of GEO comes down to one thing: content. Specifically, content on your own website, and content on the other pages across the web that AI is actually citing when it decides what to recommend.
This is worth stating plainly, because it's the piece that seems to get lost most often in how GEO is being sold and talked about right now. If a campaign isn't fundamentally focused on creating or editing content that AI will actually cite when recommending your brand, then whatever it's being called, it isn't really GEO.
The Distraction: Technical Integration
To be fair, there's a genuine category of interesting, worthwhile technical work you can do on your website that has nothing to do with content. Integrating AI directly into your site. Structuring things so AI agents can navigate and use your website more efficiently, the kind of groundwork that becomes genuinely important as agents start visiting websites to complete tasks on behalf of users, rather than just search engines indexing pages for humans.
That work matters. It's forward-looking, and it positions a business well for where the web is clearly heading.
But it isn't GEO. None of it is going to get your brand cited more often by ChatGPT, Gemini, or any other AI system, and none of it is going to improve how often you show up in AI search results. Those two things run on entirely different mechanisms, and conflating them is exactly how a lot of businesses end up paying for the wrong work.
Why Content Does the Heavy Lifting
The reason content carries 90% of the weight comes back to how AI search actually functions. When someone asks an AI a question, it searches, evaluates a large number of pages, and cites a small handful of them based on how clearly they answer the question and how much it trusts the source. Every part of that process, being found, being trusted, being cited, being recommended, runs through the actual words on the page. Not the site's technical architecture. Not whether it has slick AI integrations bolted on. The content itself.
That's true whether the content lives on your own website or on a third-party page that happens to mention your brand, a review site, a forum thread, an industry publication. AI doesn't care how technically sophisticated the site hosting that content is. It cares whether the content clearly, credibly answers the question being asked.
What This Means Practically
If you're evaluating a GEO strategy, whether it's one you're building yourself or one an agency is proposing, the test is simple: is the bulk of the actual work focused on creating and refining content, on your site and elsewhere, that's genuinely built to be cited? If most of the proposed work is technical implementation on your own website instead, you're likely paying for something adjacent to GEO rather than GEO itself.
Both kinds of work have their place. But only one of them is actually going to move the needle on whether AI recommends your brand. Know which one you're paying for.
Related reading
3 Warning Signs Your GEO Strategy Isn’t Delivering
Three warning signs to watch for when evaluating a GEO service: disconnected SEO, missing AI citation tracking, and an approach focused only on technical website changes.
Read articleStrategyAI Search Has Opened Up a Whole New Category of Entry Points
AI search creates discovery moments traditional SEO never could. Here is how to find and use these new entry points for your business.
Read articleStrategyGoogle's AI Guide Also Busts the Biggest GEO Myths Going Around
No llms.txt, no markdown copies, no special chunking, and structured data won't help you rank in AI features. Here's what Google's guide actually clears up.
Read article