AI search moves fast, and the last few weeks have moved faster than most. Between June and early July 2026, five separate data releases and product announcements, from a major referral-traffic study to three Google updates, reshaped how brands should think about AI visibility. Here’s a round-up of the stories that matter, and what to do about each one.

For global brands running paid media, SEO, or content programmes across multiple markets, this isn’t a future problem. In several categories, AI answers are already a primary discovery layer, sitting alongside (and sometimes ahead of) the traditional search results page. Here’s what changed, and what to do about it.

ChatGPT now owns AI referral traffic, but it’s sending visitors to the wrong page

A July study from Previsible, published via Search Engine Land, analysed 6.77 million LLM-driven sessions across 166 websites over 19 months. The headline number: ChatGPT now commands 92.4% of all standalone AI referral traffic, up from roughly 84% in December 2025. Monthly LLM sessions grew 9.9x to 644,478 in May 2026 alone.

Underneath that dominance, the field has flipped:

• Claude quietly overtook Perplexity in March 2026, growing 64x since late 2024, driven largely by agentic tools and enterprise adoption.
• Perplexity has fallen 61% from its traffic peak, despite launching its Comet browser globally.
• Copilot has collapsed 96% from its 2025 high.

The more actionable finding, though, is where that traffic lands. ChatGPT sends 28.8% of its referrals to internal search results pages rather than the page that actually answers the query. The model trusts a domain enough to send a visitor, but can’t always identify the right page, so it defaults to the site’s own search box. In practice, that means internal site search is no longer a housekeeping feature; it’s an acquisition surface that most sites haven’t optimised at all.

AI visibility remains highly volatile

AI visibility is still far from settled. SEMrush’s expanded 2026 AI Visibility Index grew from a 2,500-prompt pilot to an analysis of 126 million U.S. AI search prompts across ChatGPT, Gemini, Google AI Overviews, and Perplexity between January and April 2026.

The findings reveal a rapidly changing landscape:

• Only 36 brands rank in the top 100 across all four AI platforms, highlighting how difficult it is to maintain consistent visibility.
• Visibility varies significantly by industry. News & Media and Consumer Electronics are already dominated by a small group of brands, while Finance and Industrial sectors remain much more competitive, creating a valuable opportunity for challenger brands to establish visibility before these categories become more concentrated.

Google is rebuilding search and paid media around AI

A new intelligent search box, powered by Gemini 3.5 Flash

At Google I/O, Google announced what it called the biggest change to Search in over 25 years: the platform is being rebuilt with Gemini 3.5 Flash as the default model for AI Mode globally. AI Overviews now serves 2.5 billion users monthly, and AI Mode passed 1 billion users in its first year. Users can attach images, documents, videos, and open browser tabs directly to queries, reinforcing that being cited inside an AI-generated response is now a visibility metric in its own right, not just a nice-to-have alongside classic rankings.

Google published its first official GEO optimisation guide

On 15 May, Google published its first consolidated guide to optimising for generative AI features in Search. Its central message: there is no separate AI search strategy. GEO and AEO are foundational SEO applied to AI surfaces. The guide identifies five content types that reliably earn AI citation, all built on uniqueness:

• Original research
• Expert analysis
• First-hand experience
• Unique data
• Proprietary perspective

It also directly debunks a popular myth: llms.txt files receive no special treatment from Google’s crawlers.

Search Console now reports on AI visibility directly

On 3 June, Google launched dedicated Generative AI Performance Reports inside Search Console, the first native measurement of AI Overviews and AI Mode impressions. Reports cover impressions, pages, countries, and devices, from hourly through monthly granularity, plus a content opt-out toggle (effective 17 June) that removes content from AI features without touching organic rankings. Click and CTR data isn’t included yet, but impression trends by page are, for the first time, a native signal rather than something brands had to infer from third-party tools.

Paid media is changing shape too

Google is quietly testing AI-generated summaries displayed beneath sponsored ad descriptions in Search results, each carrying a disclaimer that the summary is generated independently of the advertiser, meaning Google, not the brand, decides what gets highlighted under a paid ad. Separately, Google Ads delayed the Dynamic Search Ads auto-upgrade to AI Max from September 2026 to February 2027, while automatically created assets and campaign-level broad match still migrate on schedule. From 15 June, the ad_storage setting inside Consent Mode became the sole control over Ads cookie collection, decoupled from Analytics’ Google Signals. Accounts with ad_storage denied now receive cookieless pings only, which can shrink remarketing list sizes without warning.

What global brands should do now

Taken together, these updates point to a clear set of priorities for the second half of 2026:

• Optimise for ChatGPT first, then expand. It’s the platform sending the most standalone referral volume by a wide margin, but fix your internal search UX so that traffic doesn’t bounce once it lands.
• Run a category-level AI visibility audit. Don’t assume you’re behind everywhere; concentration varies enormously by vertical, and some categories are still genuinely open.
• Treat Google’s GEO guidance as a useful benchmark rather than a definitive checklist. Use it to review your highest-traffic pages, while recognising that it reflects what Google values and doesn’t necessarily align with how other AI platforms evaluate and surface content.
• Enable Search Console’s Generative AI Performance Reports as soon as they’re available for your property and start tracking AI citation rates alongside traditional rankings.
• Audit Consent Mode configuration now. Remarketing list sizes may already be shrinking as a result of the ad_storage change.

How Adapt helps brands stay visible as AI search evolves

This is exactly the kind of shift Adapt’s global clients are navigating right now. Our AI SEO service builds the research, content, and structural signals AI systems rely on to cite a brand, across AI Overviews, AI Mode, ChatGPT, and Perplexity, market by market, without losing the cultural nuance that makes content credible in each region. Where paid media and AI-generated ad summaries are converging, our International Paid Media team keeps campaigns aligned to Google’s shifting automation timelines, and our Analytics & Measurement service helps brands close the exact measurement gap that’s leaving 45% of marketing leaders unable to track their AI visibility today.

If your team is still measuring success by blue-link rankings alone, now is the moment to close that gap. Speak with an Adapt search specialist to benchmark where your brand stands across AI Overviews, AI Mode, ChatGPT, and Perplexity, and build a plan for the categories still open for the taking.

Frequently Asked Questions