The shifting landscape of search

For years, search engine optimization (SEO) revolved around a simple premise: rank high in Google, Yahoo, or Bing, and traffic would follow. That premise is no longer absolute. The rise of generative AI platforms such as ChatGPT, Perplexity, and Google Gemini has introduced a new layer between users and websites. These AI tools now answer questions directly, pulling information from web pages and presenting it within their own interfaces. As a result, even if your site ranks well in traditional search results, you may see a decline in referral traffic as users get answers without clicking through.

This shift does not mean SEO is dead. Instead, it means the rules of visibility have evolved. Understanding how AI models select and cite sources can help you maintain—and even grow—your brand's exposure in this new environment. Below, we explore the current state of AI citations, how to measure them, and specific strategies to increase your chances of being referenced.

How AI has reshaped traffic patterns

According to an analysis by Semrush of over 50,000 websites, AI-generated traffic grew by 66% in 2025 compared to the previous year. Yet the absolute volume remains tiny—less than 0.15% of total website visits. This tells us two things. First, AI is indeed siphoning away a meaningful share of user attention. Second, most of that attention never translates into a click. Users stay inside the AI platform, consuming the answer without visiting the original source. This is a double-edged sword: your content gets exposure—your brand appears in front of potential customers—but you lose the direct traffic that traditional SEO delivered.

The key takeaway is that citation visibility matters even without a click. When an AI tool repeatedly references your website, users begin to recognize your brand as an authority. Over time, that recognition can lead to direct visits, social media mentions, and other forms of engagement. The challenge is to ensure that AI models find and prioritize your content.

Why traditional SEO tactics don't always work for AI

Traditional search engines evaluate factors like domain authority, backlinks, keyword density, and site structure. AI models, particularly large language models (LLMs), approach source selection differently. They prioritize passages that directly and succinctly answer a user's query, often favoring text that is self-contained and fact-supported. Vague or overly contextual language can be deprioritized. This is not to say that quality content and authority are irrelevant—they remain crucial—but the way they are applied by AI differs.

For example, an AI model might skip a well‑backed article that buries its key point in a long introduction, while citing a shorter piece that states the answer clearly in the first sentence. Additionally, AI models often rely on a mix of sources: official documentation, reputable news sites, user forums like Reddit, and even YouTube transcripts. Building a multisurface brand presence therefore becomes part of the optimization process.

Step-by-step: How to check if AI tools cite you

You don't need expensive software to begin assessing your AI citation status. Here are several methods, ranging from quick manual checks to deeper analytics.

Manual query testing

Open ChatGPT, Perplexity, or Google AI Mode and type 10 to 15 questions that your content should ideally answer. For each query, note which websites are cited and how often your own site appears. This zero‑cost approach gives you an immediate sense of your competitive standing. Repeat the exercise periodically to track changes.

Free tools for quick analysis

Several free online tools can provide a snapshot of your AI search presence. HubSpot's AEO Grader is one example. It analyzes your domain and returns a score along with recommendations. While not exhaustive, such tools are useful for initial diagnostics.

Google Analytics referral filtering

If you have Google Analytics set up, you can filter your traffic sources by referring domains such as chatgpt.com, perplexity.ai, and gemini.google.com. Monitor these numbers over several months to see whether your AI referral traffic is increasing or declining. This data complements manual testing by giving you real numbers on actual clicks from AI platforms.

Paid analytics platforms

Once you outgrow free methods, consider investing in a dedicated AEO (AI Engine Optimization) analytics platform. Semrush One and Otterly.AI offer plans starting around $29 per month. These tools provide detailed reports on which AI models are citing you, for which queries, and how your citation frequency changes over time.

What actually earns an AI citation

A study published at the 2024 KDD Conference, conducted by researchers from Princeton, Georgia Tech, and the Allen Institute for AI, tested nine content visibility tactics across 10,000 search queries. They achieved improvements in citation frequency of up to 40%. Here are the most effective strategies they identified:

• Allow AI crawlers: Many content management systems block crawlers like GPTBot, OAI‑SearchBot, PerplexityBot, ClaudeBot, and Google‑Extended by default. Check your robots.txt file and ensure these crawlers are permitted. This simple step prevents your content from being invisible to AI tools.
• Lead with clear, self‑contained statements: AI models extract passages rather than entire pages. When starting a new section, write an opening sentence that independently states the main point. Add nuance and context afterward. This structure makes it easy for LLMs to pull the right passage.
• Cite numbers and sources: AI models favor content that includes specific data points, statistics, and references. Vague assertions are deprioritized. Whenever possible, support your claims with credible sources.
• Create an llms.txt file: Proposed by Jeremy Howard of Answer.AI, a simple llms.txt file placed in the root directory of your site can signal to AI crawlers which pages to prioritize. While this technique has not been officially adopted by major AI companies, it is a low‑cost experiment worth trying.
• Build a brand presence across platforms: LLMs often blend information from multiple sources. A strong presence on YouTube, Reddit, LinkedIn, or other platforms increases the chances that your brand will be cited even when users ask queries unrelated to your primary website.

Long‑term strategy: Blending SEO and AEO

The fundamental principle of creating useful, accurate, and informative content has not changed. What has changed is the delivery mechanism. Traditional SEO still matters—backlinks, site speed, and clear structure remain important for ranking in Google and Bing. However, AI‑specific optimizations now need to sit alongside traditional efforts.

One practical approach is to audit your existing content with AI citations in mind. Identify pages that answer high‑value questions and rewrite their opening paragraphs to be more direct. Add relevant statistics and references where possible. Ensure your robots.txt and sitemap are configured to allow AI crawlers. Consider creating an llms.txt file as an experiment.

Another important step is to diversify your content formats. AI models increasingly incorporate video transcripts and forum discussions. Publishing explainer videos on YouTube with clear, searchable titles and descriptions can yield additional citation opportunities. Engaging in niche communities on Reddit or Stack Overflow where your expertise is relevant can also boost your brand's visibility across AI platforms.

Monitoring your progress is essential. Use the manual and free tools described earlier to establish a baseline, then revisit every few months. If you invest in a paid platform like Semrush One, you can set up alerts for when your citation frequency drops or when a competitor gains ground.

The rise of AI search is not a death knell for website traffic. It is a shift in how users discover and interact with content. By adapting your strategy to accommodate both traditional search engines and AI models, you can maintain—and even expand—your digital footprint. The tactics outlined here require relatively small changes to your content workflow, yet they can produce outsized results in citation frequency. Start by allowing the right crawlers, then refine your writing style to deliver clear, fact‑based answers early in each section. Over time, these adjustments will help ensure that when an AI model looks for an answer, it finds you first.

Source:ZDNET News