AI Now Shapes 60% of Search Journeys, Reshaping Brand Visibility

AI has become the default layer in consumer search

Six in ten search journeys now involve artificial intelligence in some form, according to new research from digital marketing agency Brainlabs. That figure represents a dramatic shift: just one year ago, only 12% of consumers reported using AI platforms for product research. Today, that number has jumped to 30%.

The remaining 30 percentage points come from passive AI exposure. Roughly 40% of Google results now include an AI Overview, meaning users read AI-generated summaries whether they chose to or not. Brainlabs projects the combined figure will reach 80% within twelve months based on current growth rates.

The research, first reported by Brainlabs and distributed by Stacker, identifies three consumer segments. Traditionalists—69% of users—stick exclusively to established search engines. Augmenters, the fastest-growing group at 30%, layer AI platforms like ChatGPT or Gemini onto their Google searches for complex queries. Dissenters, fewer than 1%, have abandoned traditional search entirely.

Why it matters

Brands optimized exclusively for Google's top-ten rankings are missing the majority of AI-driven citations. Gemini cites pages from Google's top ten results only 15% of the time, despite being a Google product. That overlap has collapsed from 76% for AI Overviews just months ago to roughly half that figure in 2026. Traditional SEO strategies no longer guarantee visibility where consumers are actually getting answers.

How large language models choose what to cite

When an LLM like Gemini receives a product research query, it generates dozens of micro-questions from the original prompt and searches Google's index across the top 100 results and beyond—not just the top ten. The system prioritizes direct answer blocks, headings that match micro-questions exactly, hard statistical data, and recent last-updated dates.

Pages that answer specific micro-questions directly can outrank higher-authority sites that bury answers in editorial content. Brainlabs client work has identified three optimization approaches that consistently increase citations: building FAQ sections around the micro-questions LLMs generate, running embedding similarity analyses to match content patterns already being cited (producing an average 140% citation increase), and treating monthly content refreshes as a technical requirement for time-sensitive topics.

The measurement challenge

No major AI platform provides first-party performance data at scale. Google Search Console's Generative AI report shows impressions but not the queries or clicks needed for optimization decisions. Third-party tracking tools exist—more than thirty at last count—but the data is unstable. Only 23% of citations remain active after fourteen days. Platforms agree on which brand to recommend first just 45% of the time. Fewer than 5% of query sets show perfect consensus across all major models.

Brainlabs recommends running 100 related prompts per category four times each, tracking AI referral traffic in Google Analytics, and documenting methodology rigorously since any AI traffic estimate will face scrutiny from leadership.

Agentic search represents the next shift

Google CEO Sundar Pichai has described plans to transform search into an "agent manager" where AI agents research, filter, and facilitate transactions without users leaving the platform. ChatGPT already offers Instant Checkout. Google has launched its Universal Commerce Protocol.

Brands should restructure content for machine consumption with explicit conclusions and clean HTML, treat proprietary data as an asset agents will want API access to, and consider building agentic platforms around genuinely valuable expert content rather than only feeding existing platforms.

The overlap between traditional SEO and AI optimization is shrinking. Brainlabs advises starting with AI referral tracking in analytics, auditing priority content for freshness dates and direct-answer formatting, and building divergent capabilities now rather than waiting for the market to mature further.

The research was conducted by Brainlabs and originally published through Stacker Media.