AI Search Stats for Marketers

A regularly updated resource of data on how brands show up in AI search, for marketers benchmarking their AI visibility and journalists looking for citable research. All statistics are drawn from Evertune's proprietary platform.

Updated August 2026 · Sourced from the Evertune Platform

Key findings

A summary of what Evertune's research has found about AI model behavior, shopping and source citations as of July 2026.

• ChatGPT shows an ad in 1 of 8 responses (July 2026)
• 50% of ChatGPT citations are listicles, and 58% of those are ranked lists, based on analysis of 40,000+ heavily cited URLs across nearly 50 product categories (April 2026).
• The typical ChatGPT-cited page has 941 words, 4 H2s, 2 H3s, 15 external links and 10 images, measured across thousands of cited pages (April, 2026)
• YouTube's citation share has roughly tripled on Google AI Mode and AI Overview since October 2025, based on ongoing citation tracking.
• 32% of all AI-cited domains are earned media, ranging from as low as 7% to as high as 87% depending on the topic.
• GPT-5.4 mini recommends 37% fewer brands than GPT-5 mini, based on ~220,000 prompts across 9 product categories (March, 2026).
• ChatGPT draws on training data 62% of the time and Gemini 54%, running live web searches the rest of the time (September, 2025).
• Shopping experiences appear in 87% of ChatGPT responses to product questions, up from 20% in October 2025, a fourfold rise in under five months (March, 2026).
• The same questions generate 838 unique URLs on Perplexity versus 6,687 on Google AI Mode, a nearly 8× difference in source diversity (2026).
• Reddit accounts for roughly 1 in 5 Perplexity citations, while even Wikipedia, ChatGPT's most-cited domain, tops out at about 5% (2026).

AI usage: monthly active users by model

Aggregated from across the web.

• AI Overviews — 2.5 billion (June 2026)
• AI Mode — 1 billion (June 2026)
• ChatGPT — 1 billion (May 2026)
• Gemini — 900 million (May 2026)
• Claude — 56 million (Q2 2026)

AI model behavior

37%

GPT-5.4 mini recommends 37% fewer brands than its predecessor.

Across ~220,000 prompts in 9 product categories, the newer model universally returned fewer brand recommendations. The top three-to-five brands were mostly stable, but brands outside that cohort saw markedly different visibility. Evertune Research, March 2026.

62%

ChatGPT relies on base model training knowledge 62% of the time. Gemini, 54%.

The remainder of the time, both execute live searches. When AI searches live, those sources become your roadmap, showing where AI looks for answers about your industry. Evertune Research, September 2025.

Varies

AI models apply very different standards of safety and recency.

Tested on Spring Break travel in March 2026, Gemini warned about specific cartel violence, Copilot recommended destinations with no caveats, and Perplexity added only generic disclaimers, making cross-model tracking essential. Evertune Research, March 2026.

AI shopping & commerce

87%

Shopping experiences now appear in 87% of ChatGPT responses to product questions.

In October 2025, ChatGPT's shopping widget appeared in 20% or fewer responses. By February 2026 that surged to 80%+ for most categories, a fourfold increase in under five months. Evertune Research, March 2026.

AI citations & sources

50%

50% of ChatGPT citations are listicles. Of that, 58% are ranked lists.

Evertune reviewed more than 40,000 URLs ChatGPT cited heavily over 60 days across nearly 50 product categories. The top domains for listicles are statista.com, forbes.com and en.wikipedia.org. Evertune Research, April 2026.

3×

YouTube's share of citations has roughly tripled on Google AI Mode and AI Overview since October.

A signal for brands investing in video content: YouTube is gaining citation share on both Perplexity and Google AI Overview. Evertune Research, 2026.