Search is changing. Users are still searching for information, but the way they find it has shifted significantly. Where people once scrolled through a list of links, they now get direct answers from AI. Google AI Overviews, AI Mode, ChatGPT, Perplexity, and Gemini all provide answers without requiring users to click anywhere. Traditional SEO, where the goal was to rank at the top of search results, is not enough on its own anymore.
AI optimization, also called GEO (Generative Engine Optimization), is the practice of making your brand and content visible inside AI-generated answers. This article explains what has changed, why it matters, and how to start working with it.
How search behavior has changed
The shift toward AI-generated answers is not gradual. It is already happening at scale.
A Growth Memo usability study of 250 AI Mode sessions found a 100% zero-click rate except for transactional queries (Indig & Johnson, 2025). Semrush data shows that 92–94% of AI Mode sessions end without an external click, compared to 35–46% for regular Google Search (Harsel, 2025). An Ahrefs study of 300,000 keywords found that AI Overviews alone reduce clicks to the number one organic result by 58% (Law & Guan, 2026).
This is what Kevin Indig, Organic Growth Advisor and author of the Growth Memo, describes as a fundamental decoupling of visibility and traffic (Indig, 2026). Users start their search inside AI interfaces and stay there. The research happens inside the AI environment. The recommendation is made there too. The click, if it happens at all, comes later and less often.
For brands, this changes what “being visible” means. Ranking number one in Google is still valuable, but it no longer guarantees that users will see you when they are making a decision. AI surfaces only a handful of brands per answer, typically around 2–4 names. If yours is not one of them, the user may never encounter you.
Zero-click is not zero value. Even when users do not click, the brands mentioned in AI answers build recognition and association over time. But that only applies to the brands that are mentioned. The ones that are not included are simply absent from the conversation.
How AI systems decide what to include
AI systems use two sources of information: what they learned during training, and what they find in real time through search.
Training data is built from content that existed before the model was trained. The more consistently your brand appeared in quality sources before that cutoff, the higher the chance it is part of what the model has learned. Real-time retrieval works differently. When a user asks a question, many AI systems run a search in the background and pull fresh information to supplement what they already know.
This distinction was a central part of my “Feed the Machine” presentation at the Baltic-Nordic SEO Summit 2026.
If your brand is absent from both of those sources, it will not appear in the answer. Being on your own website is not sufficient. AI is also not drawing from the same sources as traditional Google rankings. Research found that only 12% of URLs cited by ChatGPT, Perplexity, and Copilot rank in Google’s top 10 for the same query, and 80% do not appear in the top 100 Google results at all (Guan, 2025).
AI also needs clarity, not just presence. If the information available about your brand is vague, inconsistent, or generic, AI has little to work with. Brands that are mentioned clearly and consistently, with specific descriptions, use cases, and context, are the ones that get recommended.
Two areas to work on
AI optimization has two main areas to work on. They are complementary, and they work together.
Optimizing your own website content
The structure of your content affects whether AI can extract and cite it. Pages that answer questions directly, use clear headings, include FAQ sections, and place the main answer in the first 50 words are more likely to be used as AI sources than pages that bury information in long introductory text.
This does not require rebuilding everything. For most sites, the more useful question is not what new content to create, but what structural changes would make existing content easier for AI to extract. Small adjustments to how pages are written often have more impact than creating new pages.
Full AI content optimization guide: How to structure your content to be cited in AI search
Brand mentions on third-party sites
93.5% of AI citations come from sites you do not own (Dahlin, 2026). That means the signals AI uses to understand and recommend your brand are largely coming from external sources: articles, review platforms, comparison pages, industry publications, and community discussions.
Brand mentions are different from citations. A citation is when AI references your page directly with a link. A brand mention is when your brand is named and described in a third-party source, building the associations that AI draws on when forming an answer. Both matter, and they serve different functions.
To influence what AI says about your brand, you need to be present in the sources AI already trusts. That means placing specific, positioned descriptions of your brand on pages that rank for your keywords, on domains that are already cited in AI Overviews, and on platforms like Reddit and YouTube where AI systems actively look for signals. Branded web mentions, branded anchors, and YouTube mentions are among the signals that correlate most strongly with AI visibility in current research.
Knowing where to place brand mentions is one thing. Having a repeatable process for doing it systematically is what makes the difference over time. Working with the WhitePress platform, I have tested and documented a process for identifying where AI is currently looking for answers in your category, what is being said about your brand and your competitors, and how to place brand mentions that AI can actually use.
How to get your brand visible in AI search: AI optimization and brand mentions.
The complete LLM-feeding process: Feed the Machine: A Guide to Off-Page LLM Optimization
My opinion on where to focus
The right starting point depends on where your brand stands today.
If you are a well-known brand with strong authority, a high branded search volume, and a clear category presence, your own website content is often the higher priority. AI is already likely to know who you are. The question then becomes whether your pages are structured in a way that lets AI extract and cite specific answers about your products, services, or expertise. Optimizing existing content for clarity and structure will do more than trying to build external presence that is already there.
If you are a smaller or less established brand, if your competitors have more visibility than you, or if the market you operate in is competitive, third-party brand mentions matter more. AI forms its recommendations from what it finds across the web. If your brand is not being mentioned in the sources AI trusts, it will not surface in answers, regardless of how well your own site is structured. In that case, working on brand message consistency and building presence on external sources is where to start.
A good way to decide is to check your current AI visibility. Search for the questions your potential customers are asking in ChatGPT, Perplexity, and Google AI Mode. See whether your brand appears, what is being said, and which competitors are showing up instead. That will tell you more about where the gap actually is than any framework.
In most cases, working with both is the right approach. They reinforce each other. But if you have to choose where to start, the answer depends on how visible your brand already is.
Why starting now matters
AI systems are not static. They update their training data, adjust their retrieval preferences, and shift which sources they treat as authoritative. The associations AI has learned about brands now will influence what it says about them for months or years.
That is not a reason to panic. It is a reason to start building these signals before your competitors do. Brand visibility in AI answers compounds in the same way that organic rankings compound in traditional SEO. Earlier work creates an advantage that becomes harder to close over time.
Traditional SEO still matters. Google rankings, page structure, and content quality all feed into both traditional search and AI visibility. The difference is that AI optimization adds a layer of off-page signals, brand consistency, and source placement that traditional SEO alone does not address.
The goal is not to replace one approach with another. It is to work with both, because users are now searching in both places.
References
Dahlin, K. (2026, March). Feed the machine: A guide to off-page LLM optimization. WhitePress. https://www.whitepress.com/en/knowledge-base/6189/feed-the-machine-a-guide-to-off-page-llm-optimization
Guan, X. (2025, September 3). Only 12% of AI cited URLs rank in Google’s top 10 for the original prompt. Ahrefs. https://ahrefs.com/blog/ai-search-overlap/
Harsel, L. (2025, September). AI Mode zero-click rate [Data]. Semrush. As cited in position.digital (2026). 100+ AI SEO statistics for 2026. https://www.position.digital/blog/ai-seo-statistics/
Indig, K. (2026, January 19). The great decoupling. Growth Memo. https://www.growth-memo.com/p/the-great-decoupling
Indig, K., & Johnson, A. (2025, October 6). What our AI Mode user behavior study reveals about the future of search. Growth Memo. https://www.growth-memo.com/p/what-our-ai-mode-user-behavior-study
Law, R., & Guan, X. (2026, February 4). Update: AI Overviews reduce clicks by 58%. Ahrefs. https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/