AI, ChatGPT & Google AI Overviews: How AI SEO will optimize search behavior in 2025

How is AI SEO changing the rules of the game in marketing? In an interview with performance marketer Severin Lucks from the agency DELUCKS, you will learn why classic SEO strategies are reaching their limits, how Google AI Overviews, zero click, and search queries via tools such as Gemini influence relevant traffic—and what opportunities AI SEO now opens up.

AI 2025: AI SEO as a game changer in digital marketing

Severin, before we dive into the topic, please introduce yourself and your agency. What exactly do you do—and for whom? What role does artificial intelligence play in this?

Thank you very much for allowing me to be here. We are a small performance marketing agency based in Munich whose mission is to eliminate superficial knowledge among SMEs and start-ups and to turn their gaze from the outside to the inside. Because what many companies in Germany unfortunately really don’t have is the ability to develop truly good marketing measures and content that stick in the minds of potential customers and employees.

Instead, poor prompts and ChatGPT are used to generate an abundance of interchangeable content that merely reflects the company’s perspective. However, this leads to poor results, low conversion rates, and dependence on expensive advertising, which is extremely short-sighted.

Many people still think of keywords and meta texts when it comes to SEO. In your opinion, what are three key changes that companies should be aware of with regard to AI SEO and generative engine optimization?

Keyword-optimized metadata and content are still building blocks for SEO. However, this alone will not lead to success. Without truly good, unique content, you won’t get anywhere these days, as there is currently AI content inflation. My three main points are therefore:

1. Establish authority

The primary goal of AI SEO today is to build authority, i.e., to become as visible as possible on certain topics in order to give search engines or LLMs a reason to mention your company. So it’s not just about optimizing individual pieces of content with keywords and metadata, but about creating a range of high-quality, unique content that really interests potential customers and employees and makes them stop and take notice.

2. Build your reputation

The next step is to build up your reputation with this content and your own offerings. This means not engaging in paid link building, but rather doing really good PR, online reviews, and social media marketing to increase your reach as naturally as possible. However, if you don’t yet have the “street credibility,” you’ll still have to dip into your advertising budget and, for example, amplify content, place advertorials, or purchase user-generated content.

3. User Experience verbessern

The third component is user experience: Google also measures so-called user signals as indirect ranking factors, i.e., dwell time and bounce rate, and the conversion rate also has an indirect effect on SEO success. Therefore, the information on your own website should be presented as attractively as possible—also to avoid boring or overwhelming visitors. This can be achieved with animations and videos, but also with interactive elements such as configurators.

When you start working with a new client, what do you look at first when it comes to content? What constitutes “good content in 2025” for you—also in the context of AI SEO and AI-supported search?

In addition to traditional content audits, we check whether the content has already been referenced by AIs and what the reactions are on social media and in newsletters in order to assess the status quo. Presence in AIs can be measured relatively easily with Google Analytics, for example, as you can search for ChatGPT or Perplexity in the traffic sources and view the referenced content.

As I mentioned earlier, the relevance of the content is extremely important. That’s why we review the content afterwards. Today, this no longer has to be done manually. Instead, we can use AI to evaluate whether the content is relevant to the topic, unique, and gets to the point quickly.

Beyond that, it’s also important that people critically review the content, or at least the results of the audits, and think about how to create additional value. After all, we want to develop content that’s worth pausing for and that sticks in people’s minds. AI isn’t creative enough for that, though; it’s just based on collected knowledge, i.e., machine learning.

Google AI Overviews and Zero Click Searches optimize SEO strategy

Google is currently testing what are known as AI Overviews. Can you briefly explain what these are—and why they could be an opportunity or a challenge for companies in terms of AI SEO when it comes to increasing visibility?

With AI Overviews, Google provides answers directly in the search results, usually when searching for information. AI Overviews are not yet available, or only very rarely, for offers, products, brands, and local searches.

In the best case scenario, Google’s AI answers mean you no longer have to click through to websites. From a user’s perspective, this is of course convenient because you can get to the information you want more quickly. However, for some companies, these “zero-click searches” mean they may receive less traffic.

Statistically speaking, the loss of traffic due to AI Overviews is currently still relatively low, as the quality of the answers is sometimes still poor or too superficial.

But depending on the industry and previous SEO strategy, between 15 and 60 percent of traffic could already be lost, as measured by Forbes. Information pages and purely information-heavy SEO strategies that are based on simple glossaries, for example, are particularly affected. For most companies, however, there are usually other reasons for traffic loss, such as inferior or outdated content.

In your opinion, are there clear levers for preparing content in such a way that it is better recognized by AI systems and used for AI SEO to generate more traffic?

Yes and no. There are actually no new SEO measures for optimizing AI systems. Well-structured content is still important: this means using hierarchically formatted headings, lists, and tables, as well as structured data, such as special formatting for FAQs and user-generated content.

The best lever for AI indexing is content quality. To set yourself apart from the mainstream, you can, for example, focus on data journalism, i.e., conduct your own market research and develop studies, publish comparisons, or publish correlations. That’s what I meant at the beginning with super high-quality and unique content.

Many people in marketing use ChatGPT and other tools—sometimes for ideas, sometimes for entire texts. How do you use generative AI? And what should you pay attention to so that AI SEO isn’t diluted by generic content, which then affects search results?

I use ChatGPT and similar tools every day, and it certainly saves me 50% of my previous work related to content and code. But instead of prompting everything individually, my team and I develop our own AI tools for recurring tasks:

AI Workflows

With AI workflows, we automate entire processes, such as the development of SEO-optimized content, digital advertising materials, and psychologically optimized landing pages.

To minimize the risk of test tube results, we rely on data in addition to very detailed prompts and briefings. In other words, we integrate external tools and statistics into the AI. Of course, this is only possible to the extent that it does not contain any personal data or confidential internal information.

In your opinion, how will user search behavior change with the increasing use of LLMs such as ChatGPT—and what role does AI SEO play in this?

We are only at the beginning of a massive change. In the long term, the classic Google search for information will become a niche activity – the new standard will be for people to talk to AI systems such as ChatGPT or Perplexity.

For marketing, this means moving away from the scattergun approach and focusing on real substance. Or quality over quantity. It is not enough to just “create content.” You have to ask yourself: would an AI really recommend this? If not, it’s probably too weak.

Companies therefore need clear positioning, reliable information and, above all, content that helps, not sells. Otherwise, someone else will end up selling – via AI.

SEO will therefore continue to exist. Not only because classic Google searches are still used in many areas, at least for now, but also because “next level SEO” strengthens your presence in AI assistants. Conversely, companies that fail to be found organically will become more dependent on advertising, i.e., SEA and social ads.

Looking ahead to the next year or two, how do you think Google, AI search, and content strategy will develop—and how will AI SEO change?

Google’s AI Mode is currently being tested in the US. Similar to Copilot, it is offered in a separate tab in the search engine and is chat-based, like ChatGPT.

As a result, zero-click searches will certainly continue to increase, meaning that companies that fail to adapt will see a rise in traffic loss. I fear that some companies will find it very difficult to deliver the content quality and user experience that will be necessary to compensate for this.

If I may venture a prediction, silos must be broken down now and all stakeholders must be involved in marketing so that the foundation for a completely transparent, user-centered market can be created.

Thank you for the super exciting conversation, Severin.