GEO ≠ SEO. But Solid SEO Wins More Than Half the Battle.

"SEO is dead,"

is something you've probably heard in recent years.

Some even suggest that AI will kill websites altogether, or that tomorrow’s internet will be made up of a few Large Language Models (LLMs) playing fetch with user queries.

And for a moment, Google's mid-2026 announcement of their default AI Mode mid-2026 did look like it was going to be the final nail in the coffin.

But peer behind the curtain and you'll quickly realise that there's certain things about SEO's supposed replacement, Generative Engine Optimisation (GEO), that are suspiciously familiar...

 

Is GEO just repackaged SEO?

“To win AI search, content needs to answer common queries while anticipating follow-up questions,"
was one oft-touted advice given by GEO gurus promoting this new, AI-forward optmisation strategy.

Sounds like targeting question-based, long-tail keywords to me! A tried and true SEO strategy to make sure that your website's pitch hits customer pain points. It's also a good strategy to rank for (usually) less competitive keywords first, before moving onto bigger leagues.

To be clear, answering common queries is good practice regardless of whether you're looking at it from an SEO, GEO, or pure marketing perspective.
But this is just one of many examples of how good GEO is the same as good SEO.

 

screenshot about search volume for generative ai related terms

If SEO is dead, why is competition for GEO related terms still so fierce?

 

“Chunkable content is easier for LLMs to retrieve and cite in its answers," is another piece of advice that seems to be floating around a lot these days.

Again, haven’t we always done that for FAQ rich results?

Now, there's certainly merit to monitoring your brand's AI search visibility and giving your website the best chance it has to receive recommendations. The issue is that many of these supposed GEO strategies are nothing more than marked-up SEO packages that don't bring anything new to the table at best, or can even actively harm your website (on your hard earned dime!).

 

Where GEO & SEO overlap

You're going to hear many opinions about GEO and SEO. Some say they're exactly the same, others think they're completely different.

But we have reason to believe that GEO largely draws from SEO, with some minor distinctions. And contrary to popular belief, it might stay that way for some time, because:

LLMs rely on search results to generate answers.

This shouldn't come as a surprise if you know how LLMs amalgamate and spit out the 'average' answer across all its sources.

But this also means that LLMs are unlikely to notice you if you're not first ranking with SEO. Because you won't be among the sources they analyse in the first place!

For most businesses, worrying about choosing between GEO and SEO is a waste of time. Instead, think about how you can demonstrate EEAT (experience, expertise, authoritativeness, and trustworthiness) to convince Google, LLMs, and most importantly, your customers.

 

Do you know what you're trying to achieve with GEO?

Nearly everyone's rushing to get cited by AI, but few are stopping to think if they even should.

For one, any GEO guru who guarantees that they can get you to show up on AI search is either lying or doesn't understand how AI search works.

Second of all, LLMs are designed to keep users on the platform, which means that citations don't always translate into traffic, let alone conversions. Just think about the last time you clicked on an AI citation.

As of February 2026, Google is still accounts for 190 times more traffic and nearly five times the market share of ChatGPT.

And finally, AI generated answers still have major teething issues.

An established SEO publication Backlinko recently demonstrated how LLMs could be manipulated to cite fictitious content almost verbatim, similar to how blackhat tactics shot websites to No.1 back when Google was less crowded and had fewer guardrails.

But rest assured that companies like OpenAI and Anthropic are actively patching up these exploits. This time, with Google's blueprint on their side.

We stumbled upon another issue with AI overviews ourselves while researching "FAQ rich results", only to get an AI response telling us that Google has removed them, right above an actual set of FAQ about rich results. (this information is accurate as of 8 June 2026).

google ai saying google has removed faq rich results above faq rich results

FAQs still alive and kicking despite all the death allegations.

Now, we're not denying the first mover's advantange, and ChatGPT's 2.5 billion daily prompts are definitely nothing to sneeze at. But it's worth considering sustainable strategies that help you gain customers from both traditional and AI search without risking penalties from future updates.

Because any AI search hack loophole that you exploit today can and will be quickly patched up by LLM companies. And with Google's blueprint to follow, they'll be able to do it even quicker.

At this point, you may be wondering if it's possible to cash in on this loophole before it closes.

But before you get too excited. There's a catch.

LLMs can reference your content without ever crediting your brand name.

 

The Ghost Citation problem

Ghost citations are when AI platforms link to your webpage as a source without mentioning your brand name in its generated text.

In fact, it has a tendency to do so even while recommending a competitor.

Talk about betrayal!

llm ghost citation example retrieves information from rei but recommends brooks running shoes

Imagine having LLMs using your words to recommend a competitor.

Mentioning your brand name throughout isn't a solution either. Not only is that trickier than it sounds (and can potentially trigger Google's spam or keyword stuffing filters), it may not work at all.

A Seer Interactive study analysing 541,213 LLM responses showed that brand mentions remained at zero even after brand language was embedded retroactively.

This has to do with the order in which AI priorities citations.

  1. Brand selection. The most actively mentioned brands across social media channels are more likely to recommended by AI  when queried.
  2. Citation. AI searches for citations to back up its claims only after deciding on recommendations. Hence why citations =/= recommendations.

But don't panic or throw out your blog content strategy just yet.

As we have established earlier, LLM citations don't translate as nicely into traffic as traditional SEO.

And depending on your business model, traditional SEO alone might just be enough.

Property related businesses, for example, continue to rely heavily on organic traffic. Because who wants to make buying or renting decisions based on AI descriptions and generated images?

On the flipside, if you have reason to believe that the bulk of your customers are making decisions based on AI recommendations, then genuine social media mentions and a consistent online brand presence can help LLMs converge on a common understanding of how to represent your brand.

 

Measuring GEO vs SEO

Data from SEO tracking platforms like Semrush and Ahrefs have always been directional at best. That's because unlike Google which owns the data, third party platforms can only use samples to estimate ranks and traffic.

Of course, you can still glean plenty of useful information from them, but most small businesses can get more than enough information from free tools such as Google Analytics and Google Search Console.

So think about whether you're maximising the use of these free platforms first before committing a large sum to Semrush, Ahrefs, and the like just for their AI visibility features.

Because the sampling weakness becomes even more magnified with AI reporting, where LLM answers are non-deterministic by nature.

Yes, I see you scratching your head.

Unlike deterministic Google search results where a page either ranks for #3 or it doesn't, LLMs will give you a 10,000 different answers when prompted a 100 different times by 100 different people.

At this point, even Google hasn't figured it out. Their most recent Search Generative AI Performance feature received terrible reviews for omitting clicks and combining AI Overview and AI Mode results. (Grab your popcorn and read the comments here).

We can expect GEO reporting to grow more accurate over time. For the time being, however, it's worth looking out for these metrics as opposed to taking AI reports at face value:

  1. Recommendations. Quite possibly the most valuable form of AI citation right now.
  2. Citations. Nice to have, but won't really drive traffic unless your audience loves verifying sources.
  3. Ghost citations. Somewhat of a beige flag. Might indicate that your SEO is working well, but that online presence and customer sentiment needs work.

To conclude, stop worrying about whether you're missing the train with GEO, because establishing a technically sound website with consistent branding will get you more than halfway there. Trying to skip steps isn't only unhelpful, but might get patched over by the next LLM update anyway.

So if you want to establish a lasting brand presence that doesn't just rise to the top of Google and ChatGPT, but actually stays there, reach out to see how we can get you started.

Leave a Reply