On May 7, 2026, ChatGPT fundamentally changed how it surfaces brand links, and the traffic impact was immediate and dramatic. Instead of burying brand citations in footnotes or follow-up prompts, the platform began displaying clickable brand names directly inside answers. The result: homepage referrals surged 354.7% week-over-week, while total ChatGPT referrals increased 157.7%. This shift reveals something crucial about how AI platforms are reshaping web traffic: they are no longer just search engines. They are becoming brand discovery channels.

What Changed on May 7, and Why Does It Matter?

Before the May 7 update, ChatGPT referral traffic behaved much like traditional search. Users arrived at specific product pages or blog posts that matched their query. Only about 26% to 32% of ChatGPT referrals landed on a brand's homepage. After the update, that jumped immediately to around 60% and stayed there. This is not a subtle shift. It represents a fundamental change in user behavior and platform design.

The gap between total referral growth (157.7%) and homepage referral growth (354.7%) tells the real story. Users are now arriving at brand front doors instead of deep informational pages. This mirrors how users discover brands through advertising or word-of-mouth, not how they typically interact with search engines. ChatGPT is behaving more like a brand advertising channel than a traditional search tool.

How Is AI Referral Traffic Different From Search Traffic?

The quality of AI referral traffic is outpacing its volume. While AI-driven visits remain smaller than search or direct traffic overall, the conversion metrics tell a different story. ChatGPT referral traffic converts at 7.1%, second only to paid search at 7.8% and ahead of direct traffic, organic search, social media, email, and display advertising. Users arriving from AI platforms tend to be more qualified, more engaged, and further along in their decision-making process.

After the May 7 update, the quality signals improved further. Page views per ChatGPT-referred visit increased 24%, while average time on site rose 11%. These metrics suggest that users landing on brand homepages from ChatGPT are not bouncing immediately. They are exploring, browsing, and engaging with the brand.

Steps to Optimize Your Brand for AI Visibility

• Appear on High-Authority Platforms: ChatGPT cites Wikipedia (6.2% of citations), Reddit (5.2%), and YouTube (1.7%) most frequently. Brands that establish presence on these multi-topical, authoritative platforms increase their probability of being included in AI citation pools.
• Focus on Homepage Optimization: Since 60% of ChatGPT referrals now land on brand homepages rather than deep pages, ensure your homepage clearly communicates your brand value, product categories, and navigation paths. Homepage quality directly impacts conversion potential from AI traffic.
• Prioritize Content Quality Over Volume: AI platforms cite sources that are authoritative, high-volume, and multi-topical. Creating comprehensive, well-researched content that covers multiple angles of a topic increases the likelihood of being surfaced by AI systems.

The Bigger Picture: AI Platforms Are No Longer a Single-Player Market

ChatGPT remains the largest AI platform by raw traffic volume, but the competitive landscape is shifting rapidly. Between September 2024 and March 2026, ChatGPT's web visits grew 84%, but Google's Gemini grew roughly 9 times over the same period. Claude posted roughly 770% growth in the same timeframe, while Perplexity and Grok both gained significant traction. This is not a secondary player gradually catching up. It reflects a structural shift in how consumers are using and comparing leading large language models (LLMs).

On mobile apps, the growth is even more dramatic. ChatGPT's monthly active users grew roughly 156% between September 2024 and March 2026, but Gemini's app monthly active users grew over 1,100% in the same period. Claude's app monthly active users tripled between January and March 2026 alone, driven by new product releases. The implication is clear: any visibility strategy built around a single AI platform is already a risk.

How Often Do AI Platforms Actually Cite Sources?

The citation data reveals an important nuance. In January 2025, only 0.6% of ChatGPT answers included citations. By August 2025, that had grown to 2.8%. While these percentages are small in absolute terms, the directional trend is significant. ChatGPT is citing more sources over time, and the May 7 update accelerated this trend by making citations more visible and clickable.

When ChatGPT does cite sources, the pattern is consistent and revealing. The most-cited domains include Wikipedia, Reddit, OpenAI's own domains, YouTube, Walmart, the National Institutes of Health (NIH), Google, LinkedIn, Amazon, and eBay. These are not random. They are authoritative, multi-purpose platforms with high domain authority and broad topical coverage. For brands, this suggests that appearing on these platforms, not just on your own domain, increases visibility in AI systems.

What Does This Mean for the Future of Web Traffic?

Total AI referral visits across the web grew more than 3 times between September 2024 and September 2025. The growth was not gradual. The steepest acceleration occurred between January and May 2025, before leveling off to a sustained higher baseline through the summer. This suggests that AI referral traffic has moved from novelty to established channel.

The sites receiving the most AI referral traffic reveal where users are turning to AI for answers. Tech and information platforms dominate, but health is also a major category. NIH alone receives substantial AI-driven referrals, reflecting how often users turn to AI for medical or wellness questions. In travel, Kayak and Rome2rio lead for AI referrals. In finance, Investing.com and TradingView rank highest. In news, Reuters and The Guardian sit at the top, which ties back to the citation data: open-access publishers outperform paywalled ones in AI visibility.

The May 7 update represents a watershed moment. It signals that AI platforms are not just answer engines. They are becoming discovery channels that drive qualified traffic to brand homepages. For marketers, the question is no longer whether to optimize for AI visibility. It is how to do it effectively across multiple platforms, each with different citation patterns and user behaviors.