GEO Hype or Reality? An Honest Look at the Pros and Cons

In short: Skepticism toward GEO is justified, but the shift in user behavior toward AI-generated answers is already measurable. Those who implement the fundamentals of SEO properly are well-positioned for both scenarios.

The last Business Roundtable was a heated debate when GEO (Generative Engine Optimization) came up. The panel was by no means in agreement: Is optimization for AI search systems a meaningful evolution of SEO, or just a new trend that’s primarily a way to make good money? A look at both sides of the discussion, and my take on it.

The Skepticism: Valid Doubts

Those who consider GEO to be just hype have strong arguments on their side. Three objections carry particular weight:

• The techniques used are not new. Structured data and semantic HTML have been around for over ten years; now they’re just being marketed under a new name.
• The foundation is uncertain. AI models change rapidly, and there is currently no reliable, long-term method for measuring the success of GEO initiatives.
• There is concern that consultants and agencies, in particular, are capitalizing on this uncertainty by introducing new jargon and charging fees for it without delivering clear added value.

These points aren’t just made up. If you remain skeptical, you’re not doing anything wrong.

In Practice: What We’re Actually Seeing

At the same time, based on my own experience, user behavior is already shifting in measurable ways. In the projects I oversee—particularly in the hotel industry—I’ve observed that fewer and fewer users are coming via Google and traditional search engines. They’re increasingly asking ChatGPT, Claude, or Perplexity directly for recommendations. This isn’t a universally applicable study, but rather an observation from my own practice, which is already reflected in the numbers there.

An example from my own experience illustrates this: Imagine someone sitting on the couch in the evening and asking ChatGPT for a hotel with a spa in the mountains. Three days after a website had added its FAQ schema via JSON-LD, that exact page appeared for the first time in such responses; before that, it had never appeared there—even before the user had even landed on Google. This isn’t definitive proof. But it is a striking coincidence: coincidence or strategy?

FAQ pages and structured data aren’t just valuable for AI systems. They also improve the user experience for people looking for a quick, specific answer—and that’s not just an abstract claim to me. When I book a hotel—whether for a vacation or to attend a TYPO3 event—I always look for an FAQ section on the website, as far back as I can remember. I’m interested in questions like:

• Is there parking?
• How much does parking cost?
• What are the check-in and check-out times?

If I can’t find answers to these questions, I get annoyed every single time. To me, that’s poor usability, regardless of whether an AI is reading along or not.

My Conclusion

Both sides have a point, so it’s worth taking a level-headed look instead of taking sides. GEO isn’t magic. It’s built on fundamentals that have been around for a long time: a solid technical foundation, semantic HTML, and a proper heading structure, supplemented by structured data such as JSON-LD and the FAQ schema. Anyone who has neglected these fundamentals for years should now systematically implement them. Not because it’s a trend, but because user behavior is already demonstrably shifting.

At the same time, it’s important not to blindly follow every new buzzword or fall into a frenzy of action. Proceeding step by step, using your own website as a testing ground, and observing what actually works makes more sense than blanket panic or blanket rejection.

Anyone who wants to assess their own visibility in AI systems can currently do so as part of a limited test phase at mentionradar.de.