Is AI really destroying organic search?

Google has really shaken up the SEO world. We’re taking a look at what’s behind it. For weeks, SEO experts had only one topic on their minds: “AI is destroying organic search.” But in fact, organic rankings are being destroyed by generic, toneless AI texts — content that reflects no company voice at all. Most of the debate is primarily about Google’s AI Overview, which was gradually rolled out at the beginning of the year.

What is the AI Overview

The AI Overview is shown mainly for searched information. Instead of websites, the overview appears first. The user gets a direct answer and can click the sources mentioned to reach the website. Especially for simple questions, companies lose a lot of click-through rates because the user is not looking for deeper information and leaves Google Search again.
But for very detailed and complex topics, the AI Overview can also offer an opportunity. Those who are listed there are also more likely to receive a click.
In e-commerce, there are therefore fewer losses due to the AI Overview. What is changing there, however, is search behavior. More and more products are no longer searched for using traditional search engines, but via AI chatbot.
Ranking among the big players ChatGPT, Gemini & Perplexity is therefore becoming increasingly desirable.

Why AI buzzwords don’t add value

“The company looks forward to your application — anytime and digitally.”
What does “digital” actually mean? And what does “anytime” mean in practice? Real added value only emerges when it is explained which needs are being addressed. In this case, the company offers digital meeting solutions via Teams, Discord, Zoom or Google Meet — so applicants can participate on equal terms regardless of location. Better then: communicate the user’s need and offer the solution. Simply generated texts often shorten the actual core of the brand.
Another buzzword: “Excellent performance.” What does that mean concretely? Digital excellence, excellent presence, curtain up — ChatGPT. But that sounds generic and reproducible. No user is any wiser than before. And above all — what need is this buzzword supposed to solve?
The frequent use of buzzwords without added value can therefore have a clearly negative effect on discoverability in Google AND AI chatbots.

What changes as a result of Google’s June 2025 core update?

Google has had its longest rollout in a long time — around three weeks — a clear sign of profound algorithm changes. What’s needed are contents written for people, not machines: quality, relevance and trustworthiness (keyword E‑E‑A‑T) are weighted more heavily. In addition, the SEO community speculates that this time smaller but relevant websites may also have better chances of top rankings.

What does content look like that Google really loves?

• Graphics & statistics with alt texts: Google loves valuable visualizations.
• Step-by-step instructions: practical and useful.
• Concrete examples & new perspectives: that creates identification.
• Niche content: specific, unique, precisely aligned with the need.

In short: no run-of-the-mill stuff. Always ask yourself: “Should this AI text really be adopted 1:1, or would it be better to enrich it manually with your own wording?”

How can AI graphics be used sensibly?

Comics and AI graphics can work, but only with more sophisticated prompt engineering. Test different prompts, don’t just use the first image that comes along. The text is never 100% CI-compliant, and a trained eye for AI spots that immediately. Brand colors still cannot be implemented 1:1 with AI either. But AI graphics can be building blocks if the entire content does not necessarily consist of a single AI image.

Should stock images be replaced by AI?

Replace stock images? Yes and no — people follow people: when the team is shown, it builds trust. Until now, stock images have been the substitute for that when employees didn’t want to appear on camera. At the beginning of AI image generation, companies plastered entire websites with DALL-E. Yes, this saves licensing costs and effort, but it also means giving up part of your identity as a brand. Tools like Midjourney can now imitate a style relatively well and create consistent images.
But beware: you can’t simply feed AI with images of employees; that is not GDPR-compliant.

How extensively are companies in Germany already using AI?

Source: Destatis

• In 2024, 20% of companies in Germany (with ≥10 employees) use AI — an increase of 8 percentage points compared with 2023 (12%).
• Large companies (250+ employees): 48%
• Medium-sized companies (50–249): 28%
• Small companies (10–49): 17%
• Top technologies: text mining (48%), speech recognition (47%), NLG (34%)
• Main application areas: marketing/sales (33%), production/services (25%), administration/management (24%), finance/controlling (24%)
• Main reason for non-use: lack of knowledge

AI will no longer be dispensable; we have to prepare for that. But we do not have to lower the quality to generic and reproducible content; instead, with the help of AI, we should create real added value. It is a tool – we have to be able to handle it.