Marketing to AI: Why 44% of Users Now Prefer AI Search

Who this is for

This resource is written for CTOs, security leaders, compliance owners, platform engineers, and product teams responsible for moving agentic AI from experimentation into governed production environments.

Table of contents

1. Problem context
2. Governance pattern
3. Implementation considerations
4. Next steps

McKinsey's August 2025 survey delivered a wake-up call: 44% of AI search users now prefer it over traditional search, compared to just 31% who still prefer Google. Yet only 16% of brands systematically track their AI search performance. Even market-leading brands are entirely absent from AI-generated answers in major categories.

The opportunity gap is enormous. Organizations that optimize for AI visibility can boost citation rates by up to 40%. But here's what's counterintuitive: the tactics that worked for traditional SEO can actually hurt your AI visibility.

For two decades, digital marketing has operated on a single premise: rank on search results, earn clicks, convert visitors. This model is dissolving.

Chartbeat data shows organic Google search traffic declining 33% for publishers globally between November 2024 and November 2025. Google AI Overviews now appear in a growing share of queries—projected to reach 75% by 2028.

Yet here's the paradox: visitors arriving via AI referrals convert at dramatically higher rates. Adobe research found AI-referred traffic shows:

• 23% lower bounce rates
• 12% more page views
• 41% longer sessions

Fewer visitors arrive, but they're far more qualified. The question becomes: how do you get AI to recommend you in the first place?

Researchers at Princeton University, Georgia Tech, the Allen Institute for AI, and IIT Delhi formalized Generative Engine Optimization (GEO) in their ACM SIGKDD 2024 paper. Using a benchmark of 10,000 diverse queries, they demonstrated that specific content optimization strategies can boost source visibility by up to 40% in AI responses.

Their most critical finding? Traditional keyword optimization actually reduced AI visibility by approximately 10%.

AI systems reward different signals:

• Citations and statistics over keyword density
• Authoritative language over repetitive phrases
• Expert quotations over generic content
• Third-party validation over self-promotion

This directly challenges two decades of SEO assumptions.

AI systems use cascading trust signals across four stages: Discovery, Parsing, Embedding, and Generation. At each stage, different signals influence citation decisions.

The most surprising finding from 2025 AI visibility research: brand search volume—not backlinks—is the strongest single predictor of AI citations, with a 0.334 correlation coefficient. This means brand-building activities you thought were disconnected from search now directly impact whether AI mentions you.

Other critical factors:

• Entity consistency: Your brand name and description must match across every platform
• Review scores: Products below 3.5 stars are substantially less likely to be recommended
• Third-party presence: Brands appearing on 4+ platforms see 2.8x higher citation rates
• Structured data: Sites with complete schema markup see 44% more AI citations

McKinsey's research reveals that a brand's own website comprises only 5-10% of the sources AI references. The remaining 90-95% comes from third-party sources: reviews, industry publications, comparison sites, and earned media.

This means optimizing your website alone is radically insufficient. When ChatGPT answers "best credit cards for travel," it pulls from NerdWallet, The Points Guy, and Reddit—not issuer websites.

The implication: you need a third-party content strategy as sophisticated as your on-site strategy. Digital PR, placement in comparison articles, and cultivating reviews are no longer optional—they're primary drivers of AI visibility.

Many organizations are invisible to AI simply due to technical issues:

• JavaScript rendering: Many AI crawlers don't execute JavaScript. Content behind client-side rendering may be invisible.
• Blocked crawlers: Check your robots.txt—you may be inadvertently blocking GPTBot, ClaudeBot, or PerplexityBot.
• Missing schema: JSON-LD structured data helps AI understand your content's meaning and relationships.

A new standard is emerging: llms.txt, proposed by Answer.AI. This Markdown-formatted file in your root directory tells AI systems what your site is about and which resources matter most. Early adopters like Anthropic, Cloudflare, and Cursor have already implemented it.

The organizations winning AI visibility are investing across seven domains:

1. Audit your AI presence: Search for your brand across ChatGPT, Perplexity, Gemini, and Claude. Document where you appear—and where you don't.
2. Fix technical blockers: Ensure AI crawlers can access your content. Implement server-side rendering. Deploy structured data.
3. Build third-party presence: Invest in digital PR, comparison article placement, and review management.
4. Create citable content: Publish original research, statistics, and expert analysis that give AI something unique to reference.
5. Implement llms.txt: Create an AI-readable guide to your site's most important resources.
6. Track AI metrics: Citation frequency, share of voice, and AI-referred traffic are the new KPIs.
7. Invest in brand: Brand search volume directly predicts AI citations. Awareness campaigns are now SEO strategy.

By 2028, 75%+ of search queries will include AI-generated components. By 2030, autonomous AI agents may initiate the majority of B2B procurement research. The organizations building AI visibility now will define how their categories are presented to the next generation of buyers.

This article summarizes key findings from our comprehensive whitepaper, "Marketing to AI: Business Best Practices for Visibility, Trust, and Citation in AI-Mediated Discovery." The full whitepaper includes detailed implementation frameworks, complete research citations, and tactical checklists for enterprise GEO adoption.

Glossary and definitions