Die Ära der KI-Suche ist da: "SEO ist tot, es lebe GEO!"

[Editor's Note] As generative AI rapidly reshapes the landscape of search engines, the author of this article presents a rather provocative judgment: "SEO is dead, long live GEO." With ChatGPT's monthly active users exceeding 500 million and Google officially launching AI Mode, the author believes that traditional SEO is transitioning to GEO (Generative Engine Optimization). To this end, he not only systematically compares the differences between traditional search and AI search in terms of user behavior, ranking mechanisms, content strategies, etc., but also provides a GEO implementation guide.

In May this year, ChatGPT's monthly active users reached an astonishing 500 million.

In the same month, Google also officially launched AI Mode - a "ChatGPT-style" intelligent conversational search mode integrated into Google Search. This is a decisive moment: even Google has admitted that large language models (LLMs) will become the main interface for future searches.

Meanwhile, the SEO data of many brands are also experiencing the "crocodile effect": the number of impressions is constantly rising, but the click - through rate is continuously declining. The "credit" for all this goes to AI - of course, you can also say it's the "culprit", depending on which side you're on.

I've noticed this trend in many startups: more and more users are choosing ChatGPT or other LLM tools in the "How did you hear about our product?" (HDYHAU) survey; some customers even claim that up to 30% of their weekly traffic comes from ChatGPT.

And the speed of this transformation is truly astonishing. For example, in April 2025, 10% of Vercel's registered users came from ChatGPT, while in March it was 4.8%, and in October 2024, it was only 1%.

It's not just LLMs; traditional search itself is also changing. Most startups I've worked with still get 30 - 50% of their registered users from Google. But now, Google's search result pages are rapidly transitioning to AI Overview and AI Mode.

In other words, the future of search has undoubtedly shifted towards LLMs - and this will affect every online business globally.

Clarifying Two Common Misconceptions about GEO

First, we need to clarify two common misconceptions about GEO.

Misconception 1: GEO isn't important yet; its scale is too small

The facts are as follows:

● Google's AI Overview now appears in over 50% of searches, compared to only 25% 10 months ago.

● ChatGPT currently accounts for 3% of Google's total search traffic, and it's expected to reach 10% by the end of the year and may even catch up with Google in 5 years.

● Although Google's AI Mode is still in the testing phase, judging from the promotion speed of AI Overview, it's likely to be rapidly popularized.

For all companies that rely on online channels to acquire customers, GEO is an issue that must be taken seriously right now.

Misconception 2: GEO is no different from SEO

In terms of direction, GEO and SEO are indeed quite similar: you still need high - authority, valuable content to gain exposure in both traditional search and LLM search.

But at the implementation level, GEO and SEO are two different approaches.

(1) Difference 1: Different query methods

● Keywords in traditional search usually have only 3 - 5 characters;

● LLM search is more like a natural - language conversation, and a query often exceeds 20 words, and users will also keep asking follow - up questions.

The approach of traditional SEO is to capture "high - traffic, low - intent" keywords, such as "AI meeting recording software"; while the goal of GEO is to win citations in long - tail queries, such as "List the top 10 best AI meeting recording tools currently available and compare their prices, advantages, and disadvantages". To achieve this, you must create authoritative content that is rich in content, detailed in data, includes competitive advantage analysis, and has clear usage scenarios.

(2) Difference 2: Different result display methods

● Traditional search gives you a bunch of blue links, and you choose by clicking on them;

● LLMs will try to answer your question all at once, integrating a complete answer from multiple sources, and you may not click into any web page.

However, if users do click into your web page, the conversion rate is usually very high. Take Ahrefs as an example. Although visitors from LLMs only account for 0.5%, they contribute over 12% of registered users, with a conversion rate 12 times that of traditional search.

So the goal of GEO is: first, strive to be cited, and then optimize your website to capture this "high - intent" traffic.

(3) Difference 3: Different ranking mechanisms

SEO expert Kevin Indig found that:

● In LLM search, traditional SEO indicators (such as the number of backlinks, keyword density, and page views) have little impact on ranking;

● More importantly: the authority of the website, the comprehensiveness and readability of the content.

In summary, although GEO is similar to SEO in direction, the tactical details are completely different.

How to Determine if GEO is Worth Investing In?

If you're still unsure whether GEO is worth prioritizing, you can try asking yourself the following questions:

1. How fast are users adopting AI search?

● Determine the proportion in traditional search: Use tools like Google Keyword Planner, Ahrefs, or SEMRush to evaluate the search volume of your product keywords (assuming that half of them are affected by AI Overview);

● Determine the scale of LLM search: Use Profound's Conversation Explorer to see how often questions about your product appear in LLMs;

● Conduct direct user surveys: "Do you use ChatGPT, Google AI Mode, or Perplexity for searching?"

2. How do users find you?

Add the "How did you hear about our product?" (HDYHAU) question to the registration process. Even if you can't accurately attribute the source, you can still find out if users "heard" about you through channels like ChatGPT.

If you meet any of the following criteria, you should consider prioritizing GEO:

● Over 10% of your potential customers use AI search;

● More than 5% of your users have found you through AI search.

GEO Practical Operation Guide

Okay, if you've decided to take GEO seriously, you can refer to the following specific implementation steps.

1. Lay the foundation: Technical audit

Conduct a comprehensive GEO and SEO technical audit:

● Understand the rules of GEO and SEO;

● Optimize website speed, meta - tags, and page titles;

● Add structured data, clean up dead links, and solve content duplication problems.

2. Define your brand positioning

Mixed information will confuse both AI and users. Therefore:

● Clearly convey your positioning, including core information and brand tone;

● Ensure that every channel reflects this: your website, social media, blog, and other channels;

● Maintain consistent expression across all channels.

3. Answer users' most concerned questions

The earlier you answer questions, the more likely you are to be cited by AI:

● Collect common questions from sales, customer service, and the community;

● Answer more comprehensively and practically than anyone else;

● Anticipate new questions in advance and publish answers first.

4. Update old content

Transform old web pages into new growth drivers:

● Review old content: Update data, fix links, and add details;

● Adjust the language to make it more precise and descriptive, which is beneficial for AI understanding;

● Improve the structure, clarity, SEO tags, and internal links.

5. Create new authoritative content

Be the "best answer" on the Internet:

● Publish detailed comparison pages, practical guides, and answers to popular questions;

● Use real data, expert insights, and case studies;

● You can "tailor and reuse" existing video podcasts and other materials for content production.

6. Expand off - site influence

AI doesn't just look at your website; it also looks at your entire "online footprint":

● Research which websites LLMs most often cite to obtain information related to your business.

● Profound analyzed that the top five websites most often cited by ChatGPT and Google AI are: Wikipedia, YouTube, Reddit, Quora, and LinkedIn;

● Find ways to appear on these platforms, such as participating in forums, publishing influential content, and writing column articles.

7. Monitor, measure, and optimize

Search is a "compound - interest game":

● Track LLM visibility, brand mentions, user sentiment, click - through rates, and HDYHAU data;

● Continuously fill content gaps and conduct technical optimizations;

● Re - evaluate conversion optimization and redirect strategies to capture the middle - of - the - funnel (MOFU) conversions.

Finally, want to know the approximate time frame for GEO implementation? From my experience, from auditing the foundation, defining the positioning, optimizing content, to expanding influence and continuous iteration, it can be initially implemented within a quarter.

Remember, it's not a question of whether LLMs will affect you, but "when" they will affect you.

Original link: https://substack.gauravvohra.com/p/search-is-dead-long-live-llms-winning-in-the-era-of-llms-ai-overviews-and-geo

This article is from the WeChat official account "CSDN". Author: Gaurav Vohra; Translator: Zheng Liyuan. Republished by 36Kr with permission.