You've likely noticed that organic search traffic is quietly declining, even though your rankings haven't budged. Does this mean traditional SEO is dead? Absolutely not. But should you ignore AI search as an unchangeable black box? Also no. The rules that once governed search engine algorithms—backlinks, domain authority, performance optimization, and high-quality content—haven't vanished. However, AI search engines treat these factors very differently, creating both opportunities and challenges.

The reality is that good content is now more likely to be cited by AI tools, even if your website has a lower domain authority. At the same time, most users get their answers directly within the AI platform, without ever clicking through to your site. This means fewer direct visits but more brand exposure. As long as your content is cited by AI search engines, your audience sees your name, whether or not they register as inbound traffic.

The current state of AI search traffic

According to Semrush's analysis of over 50,000 websites, AI-driven traffic grew by 66% in 2025 compared to the previous year. Yet AI still accounts for less than 0.15% of total website visits. This statistic reveals two critical insights: first, AI is capturing a significant portion of search engine traffic, but second, most of that traffic stays within the AI platform. ChatGPT, Perplexity, and Google Gemini now serve as both the starting point and final destination for many internet users. They find everything they need without ever leaving the app.

Websites that have prioritized traditional SEO are suddenly discovering that AI tools don't rank them the same way. Search engines used to reward sites with a rich library of content and demonstrated authority. AI also values those things, but it primarily looks for text that answers a user's question precisely and immediately. This shift requires more targeted optimization than technical SEO typically provides. Along with a robots.txt file and a sitemap, you might also consider adding a dedicated llms.txt file for AI agents and bots.

How to check if AI engines are citing your site

There are several ways to assess your visibility on AI search, depending on your time and budget. The simplest method: go to ChatGPT, Perplexity, or Google AI Mode and ask 10–15 questions that your content should answer. Note who gets cited and how often. This is free and takes little time.

Many free tools like HubSpot's AEO Grader provide a quick analysis, giving you a decent idea of your brand's presence in AI search engines. If you have Google Analytics, filter referrals from chatgpt.com, perplexity.ai, gemini.google.com, etc. Track this over several months to see if your presence improves or worsens.

For deeper insights, consider paid AEO analytics platforms like Semrush One or Otterly.AI. Plans start as low as $29 per month. These tools monitor your brand mentions across multiple AI models and provide data on citation frequency, sentiment, and competitive positioning.

What actually earns an AI citation

A study published by Princeton, Georgia Tech, and the Allen Institute for AI at the 2024 KDD Conference tested nine content visibility tactics across 10,000 search queries. They achieved improvements of up to 40% in AI citations. Here's what they found works:

• Don't block AI crawlers. Many content management systems block GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot, and Google-Extended in robots.txt. Ensure these are allowed so AI tools can access your content.
• Write passage-friendly openings. AI models extract passages, not entire pages. When you start a new section, lead with a self-contained sentence that clearly states the key point before adding nuance or context.
• Cite numbers and sources. AI tools prioritize credible content. Vague statements are deprioritized during verification. Include data, statistics, and references to authoritative sources.
• Create an llms.txt file. This low-cost experiment, proposed by Answer.AI's Jeremy Howard, involves placing a simple text file at your site's root directory listing important pages for AI crawlers. While not yet officially supported by major AI companies, it's worth testing.
• Build a multi-platform brand presence. LLMs pull content from diverse sources, including YouTube videos and Reddit posts. A strong social media presence increases your chances of being cited across different platforms.

How AI search optimization differs from traditional SEO

Traditional SEO focuses on keywords, backlinks, page speed, and mobile responsiveness. AI search optimization (AEO) prioritizes information scent, answer precision, and content structure. While backlinks still matter, AI models weigh them less than the direct relevance of a passage to a user query. This means that a well-written paragraph on a low-authority site can outrank a poorly written page from a high-authority domain.

Another key difference is the importance of redundancy. In traditional SEO, you want unique content to avoid duplication penalties. For AI, redundancy can be beneficial. If the same fact appears across multiple trusted sources, the AI is more likely to include it in its response. This is why news agencies and government publications often dominate AI citations—they say the same thing in slightly different ways across many pages.

AI also favors content that is written in a neutral, factual tone. Overly promotional language or excessive opinions reduce the likelihood of citation. The Princeton study found that content with an academic or journalistic tone was cited 20% more often than content with a marketing tone.

Tools to measure and improve your AI visibility

Beyond free methods and paid platforms, several specialized tools can help you track AI citations. Semrush One includes an AI visibility report that shows your brand's presence across ChatGPT, Perplexity, and Gemini. Otterly.AI monitors citations in Gemini, Claude, and other models. These tools also provide recommendations for improving your content structure and metadata.

Google's own Search Console now includes an AI overviews report that shows impressions and clicks from AI-generated search results. This is a good baseline for measuring your organic AI performance, though it only covers Google's ecosystem.

For a DIY approach, you can use Python scripts with the OpenAI API to query ChatGPT about specific topics and extract which sources it references. This is more technical but gives you complete control over the testing process.

Practical steps to implement today

Start by auditing your robots.txt file. Remove any lines that block the crawlers mentioned earlier. Then, review your top 10 most visited pages and rewrite their opening paragraphs to be self-contained, clear, and data-rich. Include at least one statistic or source citation per section.

Next, create an llms.txt file. Open a text editor, write a brief description of your site, and list the URLs of your most important pages (e.g., your homepage, about page, and top articles). Save it as llms.txt and upload it to your site's root directory. This is a low-risk experiment that might yield significant rewards.

Finally, expand your social media presence. If your brand doesn't have a YouTube channel or Reddit presence, consider creating one. AI models increasingly pull from these platforms, especially for how-to and opinion content.

Understanding the mechanics of AI citation

When an AI model like ChatGPT generates a response, it doesn't simply retrieve the best page. Instead, it processes a mix of retrieval-augmented generation (RAG) and inherent training data. RAG systems search a vector database of indexed web pages and return the most relevant passages. This means that your content's embedding—a numerical representation of its meaning—must closely match the user's query embedding. This is why writing in clear, direct language that mirrors common search phrases is more important than ever.

AI models also use a process called perplexity filtering. If a passage is too similar to many others, it may be considered low-value. But if it offers a unique insight or a fresh angle, it stands out. This is where original research, expert interviews, and case studies can give you an edge.

The Princeton study also noted that AI models tend to favor content from .edu, .gov, and .org domains, even when their domain authority is lower than .com competitors. If your site is a .com, consider building partnerships with educational institutions or contributing guest posts on .org sites to boost your citation potential.

Future trends for AI search optimization

As AI search tools become more sophisticated, the gap between traditional SEO and AEO will narrow. Google is already integrating AI overviews into its main search results, and competitors like Perplexity are experimenting with agentic AI that can execute tasks on your behalf. By 2026, we can expect AI citations to carry more weight in brand perception, even if direct traffic remains low.

Early adopters of AEO tactics will have a first-mover advantage. Those who ignore the shift risk losing visibility as AI becomes the default interface for information. The core principles remain the same: create useful, informative content. But the way AI consumes and presents that content demands a subtle but important evolution in how we write, structure, and promote our work.

Source:ZDNET News