How Do Technical SEO Factors Impact AI Search? [Study]

Digital discovery is undergoing a meaningful shift. AI platforms like ChatGPT, Perplexity, and Google’s AI Mode are changing how people find and consume information online.

As AI-powered search experiences become more prominent, one assumption keeps coming up: that the same technical SEO factors that help pages rank in Google will automatically influence whether they’re surfaced or cited by large language models (LLMs). 

But that transfer isn’t guaranteed. Some technical signals may carry over clearly, others may matter only indirectly, and some may function more as prerequisites than ranking drivers. 

This study set out to examine which technical SEO factors show measurable correlations with AI citations, and where long-held assumptions start to blur. 

Key Takeaways: Technical SEO for AI Search

Our analysis of 5 million cited URLs reveals a clear pattern: AI platforms consistently cite pages that tend to have strong technical foundations, many of which are associated with traditional SEO success. 

These correlations don’t prove causation, but they do point to technical SEO as a foundation for AI visibility.

What correlates with AI visibility:

1. Structured data implementation: Organization, Article, and BreadcrumbList schema appear most frequently on cited pages, with higher implementation rates on pages cited by Google AI Mode
2. URL structure patterns: URLs with 17–40 character slugs receive the most citations, and cited URLs tend to use descriptive but concise paths
3. Strong user engagement signals: Top-cited pages demonstrate higher visit duration, lower bounce rates, and better conversion metrics across all traffic sources.

The Search Landscape Is Splintering

Search no longer means just Google. AI platforms like ChatGPT, Perplexity, and Microsoft Copilot are creating a parallel discovery ecosystem where users have conversations instead of scanning blue links.

Current Market Share

Traditional Search 

• Google: 90.06% (a small but steady decline) 

AI/LLM Discovery 

• ChatGPT: ~80%
• Perplexity: ~11%
• Microsoft Copilot: ~5% (growing fast from just 1%)
• Others: ~4%

For SEO teams, this fragmentation changes things. 

Traffic forecasts based solely on Google rankings now miss a growing slice of discovery. 

Google is still dominant, but it’s no longer the only place where people start their journey. AI platforms are carving out their category of discovery, and they’re growing fast. And to meet that new slice discovery, brands need visibility strategies that span both traditional search and AI platforms. 

The New KPI: AI Visibility

AI visibility measures how often—and how accurately—your brand appears in AI-generated answers. This includes benchmarking brand mentions (whether your brand is mentioned within an AI response) and citation frequency (how many times AI platforms reference your content as a source).

It’s no longer just about showing up in Google. Today’s brands are asking a new set of questions:

• “Does ChatGPT mention us when people ask about our product category?” For example, if a ChatGPT user asks “What’s the best project management software for remote marketing teams?” does Asana or Monday.com appear in the answer alongside mentioned brands?
• “Are we being cited as an authoritative source by AI platforms?”
• “What’s the quality of traffic we get from AI-generated responses?”

These aren’t fringe questions. They’re becoming central to modern digital strategy. AI visibility is now a performance metric. 

Study Methodology

We analyzed 5 million URLs cited by ChatGPT Search and Google AI Mode to identify technical SEO patterns correlated with AI visibility.

Our study collected the following data:

• 5 million cited URLs across ChatGPT Search and Google AI Mode
• 378,000 URL citations analyzed via Botpresso for structure patterns
• Engagement metrics pulled from Semrush’s Traffic Analytics database
• Schema markup analysis pulled from Semrush’s Site Audit 

With this data, we analyzed:

• Ranking position – How prominently cited URLs appeared within the list of sources for an AI response (positions 1-20)
• User engagement signals – Visit duration, bounce rate, pages per visit, and conversion rates of traffic to the URLs in the sample
• Schema markup – Presence and types of structured data on the URLs in the sample (Organization, Article, FAQ, etc.)
• URL structure – Slug length of cited URLs

Note: This study identifies correlations, not causation. However, the consistency across 5 million URLs and multiple platforms suggests these technical factors create favorable conditions for AI visibility.

Enter the “Educated Click”

One emerging pattern we observed, though not directly measured in this study, is what we call the “educated click.”

Users who land on sites from an AI engine often behave very differently. They’ve already been briefed by the AI; they know who you are and why you matter. I call this the “educated click.”

These users tend to:

• Engage more deeply
• Convert faster
• Bounce less
• Arrive with clear, informed intent.

Finding 1: User Engagement Correlates with AI Citations

URLs cited by AI platforms show a consistent pattern: they tend to have higher user engagement metrics. 

Key findings:

Top-ranked citations (in cited positions 1-5) are URLs with:

• Higher visit volume and unique visitors
• Longer session durations
• More pages per visit
• Higher conversion rates

Comparing the two LLMs, Google AI Mode had a tendency to cite pages with higher engagement than ChatGPT Search, particularly for page views and purchases per visit.

Why this matters for technical SEO: AI-cited pages tend to demonstrate stronger user engagement signals. Technical factors like fast load times, clear site structure, and mobile optimization don’t just improve user experience—they may help create conditions associated with higher AI citation rates.

It’s important to note that user engagement is measured after a click, while AI citation happens before one. This means engagement itself is unlikely to be a direct input signal for AI systems. Instead, these metrics likely act as proxies for content quality, trust, and usefulness. Pages that consistently satisfy users may also share underlying characteristics—clear structure, strong technical foundations, and credible signals—that make them more likely to be surfaced or cited by AI platforms.

Finding 2: URL Structure Patterns

Analysis of 378,000 citations across major AI platforms reveals a clear pattern in URL structure.

Key finding: URLs with slug lengths between 21-25 characters received the highest number of citations (~87,000), followed by slugs in the 6-10 character range (~57,000 citations).

URLs with moderate slug lengths (17-40 characters) consistently outperformed both very short (1-5 characters) and very long (56+ characters) slugs.

What this means: While this shows correlation rather than causation, the data suggests URLs cited more frequently by AI platforms tend to have descriptive but concise slugs. Extremely short slugs (often homepage or category pages) and overly long slugs (often deeply nested or keyword-stuffed) appear less frequently among cited URLS. 

Actionable insight: Consider URL slugs between 17-40 characters that clearly describe page content without excessive length or parameters.

Finding 3: Structured Data and AI Citations

Webmasters have started to notice that OpenAI’s crawler is busy. In fact, on many sites, it’s appearing more frequently in crawl logs on many sites. This surge creates a few implications:

• Crawl budget management – If OAI-SearchBot is consuming significant server resources, you may need to prioritize which pages it accesses through robots.txt rules or crawl-delay directives.
• Log file monitoring – Track which AI crawlers are visiting your site, how frequently, and which pages they’re prioritizing. This reveals what content AI platforms consider valuable.
• Structured data prioritization – With more AI crawlers checking your site, proper schema markup could help provide clearer contextual signals about your content.

What Structured Data Does

While its value in AI optimization is still debated, early experiments suggest structured data helps with:

• Identifying entities and relationships
• Clustering content by topic
• Enhancing citation accuracy in AI answers

With this in mind, we checked how often pages cited by AI had different types of structured data.

Here’s what the research shows: 

Pages cited by AI show a clear pattern: they’re far more likely to implement a few specific schema markup types. While this doesn’t prove schema causes citations, the correlation is strong enough to keep an eye on.

Top 3 schema markup items:

1. Organization: 25% (ChatGPT), 34% (AI Mode)
2. Article: 20% (ChatGPT), 26% (AI Mode)
3. Breadcrumb: 15% (ChatGPT), 20% (AI Mode)

Other notable schema items:

• SiteLinks_SearchBox: 5% (ChatGPT), 7.5% (AI Mode)
• FAQ: 3% (ChatGPT), 5.5% (AI Mode)
• LocalBusiness: 2% (ChatGPT), 3.5% (AI Mode)
• Product: 1.5% (ChatGPT), 2.5% (AI Mode)
• ReviewSnippet: 2% (ChatGPT), 3.5% (AI Mode)
• Video: 0.5% (ChatGPT), 1.5% (AI Mode)

Compared to ChatGPT Search, Google AI Mode consistently cites pages with higher schema implementation rates across all types, particularly Organization, FAQ, and Site Links Search Box markup.

Key finding: Open Graph and schema.org (JSON-LD) appear on the majority of cited pages across both platforms:

• Open Graph: ~60% (AI Mode) and ~40% (ChatGPT)
• Twitter Cards; ~50% (AI Mode) and ~30% (ChatGPT)
• Schema.org (JSON-LD): ~40% (AI Mode) and ~30% (ChatGPT)
• Schema.org (Microdata): ~10% (both platforms)
• Microformats: ~5% (both platforms)

For context: Open Graph and Twitter Cards are widely adopted across the web (especially on established sites), which likely influences these high percentages. 

Why this matters: You don’t need to implement every structured data format. AI platforms appear capable of understanding content regardless of whether you use Open Graph, schema.org, or both. Focus on testing formats to see what makes an impact on your visibility rather than chasing perfect coverage.

While Schema.org adoption remains relatively low, Open Graph’s widespread presence suggests AI platforms can extract entity and metadata information from multiple structured data formats.

Technical Tips for AI Visibility

Server-Side Rendering (SSR)
JavaScript-heavy sites are a challenge for AI crawlers. If your content isn’t rendered server-side, it may be more difficult for AI crawlers to index or cite.

Conversational Formatting
AI systems tend to surface content that uses Q&A-style formatting, structured summaries, and clearly organized copy. 

Structured Data 2.0
It’s not just about marking up FAQs anymore. Testing broader schema coverage may help provide clearer signals across your entities, authors, and articles built for semantic recognition, not just SEO.

Key Takeaways and Next Steps

This study suggests that technical SEO fundamentals still play an important role in AI search visibility, but not always in the ways teams might expect. Rather than acting as direct ranking signals, many technical factors appear to create the conditions that make content easier for AI systems to retrieve, interpret, and cite.

Not all signals transfer equally. User engagement metrics likely reflect underlying content quality rather than influencing AI systems directly. URL structure and structured data appear to function more as clarity and accessibility signals than as optimization levers on their own. Taken together, these findings point to a shift in how technical SEO should be evaluated for AI search: less as a checklist, and more as a foundation that enables visibility when other signals align.

For teams navigating AI-driven discovery, the takeaway isn’t to chase new technical tricks, but to test which fundamentals meaningfully support AI visibility in practice. As AI search continues to evolve, understanding how these signals interact—and where their limits lie—will be essential to building durable visibility strategies.

What to prioritize for testing:

1. Test core schema markup: Prioritize Organization, Article, and BreadcrumbList schema. Add FAQ schema for informational content and Product schema for e-commerce
2. Keep URL slugs between 17-40 characters: Optimize your URLs to clearly describe content without unnecessary parameters or nesting
3. Improve engagement signals: Focus on technical factors that drive user engagement: page speed, mobile optimization, clear site architecture, and engaging content
4. Maintain AI crawlability: Ensure AI crawlers can access and parse your content efficiently through clean HTML, proper heading hierarchy, and crawlable site structure

AI-driven discovery is still evolving, but its influence on how users find and evaluate brands is already visible. Teams that start measuring and testing now will be better positioned as these systems mature.