2026 AI Visibility Report: What Changed in AI Search From January to July

AI search visibility became a measurable marketing priority in 2026.

Between January and July, Google made Search more conversational, personalised and agentic. OpenAI expanded product discovery and advertising within ChatGPT. Microsoft gave website owners citation-level reporting. Regulators introduced enforceable controls over how publisher content is used in generative search.

At the same time, independent research found that AI-generated summaries can reduce website clicks, increase zero-click behaviour and cite sources that do not appear among the highest traditional search results.

The conclusion is not that SEO has stopped working. The change is that rankings and traffic no longer show the complete discovery journey. Brands must now understand whether they are being mentioned, cited, compared and recommended inside AI-generated answers.

What Does AI Visibility Mean in 2026?

AI visibility measures how often and how prominently a brand, product, website or expert appears in answers generated by platforms such as ChatGPT, Google AI Mode, Google AI Overviews, Gemini, Claude, Microsoft Copilot and Perplexity.

Traditional SEO primarily measures positions, impressions, clicks and conversions. AI visibility introduces a different set of questions:

• Does the answer mention the brand?
• Is the brand recommended or only referenced?
• Which webpage receives the citation?
• Which competitors appear instead?
• Is the information accurate?
• Does the AI describe the brand positively, neutrally or negatively?
• Does the brand appear consistently across repeated responses?

This distinction matters because AI-generated answers can influence a decision without producing an immediate website visit. A customer may compare providers, build a shortlist and form an opinion inside an AI conversation before clicking any link.

Which AI Search Numbers Matter Most?

The three widely shared figures from the report are supported by published sources, but each needs the correct context.

These numbers should not be presented as universal benchmarks. The 9.5 billion figure is a Similarweb traffic estimate, while the click and zero-click findings come from a specific experimental study that remains a working paper.

What Changed in Google Search During Early 2026?

Google began the year by making AI Search more personal and conversational.

On January 22, Google introduced Personal Intelligence in AI Mode for eligible Google AI Pro and Ultra subscribers in the United States. The opt-in feature could use information from Gmail and Google Photos to provide more contextual answers.

On January 27, Google made Gemini 3 the default model for AI Overviews and allowed users to continue from an AI Overview into follow-up questions. The experience moved closer to a continuous conversation rather than a single search and results page.

The marketing implication is significant. Search results can now be shaped by both public web information and the user’s personal context. Two people asking a similar question may receive answers informed by different preferences, histories and situations.

Brands therefore need more than keyword relevance. They need clear entity information, consistent product details, strong external references and content that can answer specific questions without depending on hidden context.

Source: Google: Personal Intelligence in AI Mode

How Did AI Search Become a Commercial Discovery Channel?

OpenAI made shopping and advertising a more visible part of ChatGPT during the first half of 2026.

On March 24, OpenAI launched richer product discovery experiences. Users could browse products visually, compare options side by side, refine recommendations through conversation and use images to find similar products.

OpenAI also expanded the Agentic Commerce Protocol to improve the freshness and coverage of merchant data.

On May 5, OpenAI expanded its advertising pilot. Advertisers gained access through partners and a beta self-serve Ads Manager, along with cost-per-click bidding and additional measurement tools.

OpenAI stated that advertisements would remain clearly separate from ChatGPT’s answers.

These developments created several distinct forms of commercial visibility:

A brand can succeed in one layer and remain absent from another.

For example, an article may receive a citation without the brand being recommended. A product may appear in a comparison while the company homepage receives no link.

Sources:

• OpenAI: Product discovery in ChatGPT
• OpenAI: New ways to buy ChatGPT ads

How Large Did Google’s AI Search Products Become?

Google reported that AI Mode surpassed one billion monthly active users globally by May 2026. It also stated that AI Mode query volume had more than doubled every quarter since launch.

In June, Google reported that AI Overviews had more than 2.5 billion monthly active users, while AI Mode remained above one billion monthly users.

These are company-reported usage figures and should be labelled accordingly.

Google also introduced more ways for AI responses to connect users to websites. Updates announced on May 6 included additional inline links, suggestions for further reading and website previews.

On May 27, Google brought Preferred Sources into AI Overviews and AI Mode and introduced a “Highly Cited” label intended to highlight original or influential reporting.

For publishers and brands, these changes reward content that contributes something identifiable:

• Original research
• First-hand expertise
• Clear authorship
• Accurate statistics
• Unique product information
• Strong visual evidence
• Transparent methodology
• Independent third-party references

Publishing another generic summary of information already available elsewhere gives an AI system little reason to choose that page as evidence.

Sources:

• Google: AI Mode insights
• Google: Generative AI discovery updates

Is the AI Market Still Dominated by ChatGPT?

ChatGPT remained the largest generative AI platform by web traffic, but the market became more fragmented.

Similarweb estimated that ChatGPT’s share of worldwide generative AI website visits fell from approximately 76% in June 2025 to around 53% by May 2026.

Over the same period, Gemini increased to roughly 27% to 28%, while Claude approached 9%.

This does not necessarily mean ChatGPT usage collapsed. The category expanded while competing platforms gained a larger portion of it.

For marketers, the practical lesson is simple: measuring only ChatGPT no longer represents the entire AI discovery market.

Visibility should be monitored across the platforms customers actually use, with separate benchmarks for each engine.

Source: Similarweb: Generative AI statistics

How Are AI Overviews Affecting Website Traffic?

The strongest 2026 evidence suggests that AI summaries can reduce outbound traffic for informational searches.

A randomised field experiment by Saharsh Agarwal and Ananya Sen used a browser extension to assign participants to Google Search with or without AI Overviews.

Conditional on an AI Overview appearing:

• Outbound organic clicks fell 39.8%
• Zero-click searches increased 34.5%
• Sponsored clicks were not materially affected
• Overall search frequency was not materially affected

A separate study by Mehrzad Khosravi and Hema Yoganarasimhan analysed 161,382 matched Wikipedia article-language pairs.

It estimated that exposure to AI Overviews reduced daily traffic to English Wikipedia articles by approximately 15%.

The decline varied by topic and was larger where a concise generated answer could satisfy the informational need.

These findings should not be applied equally to every industry. A short factual query may be resolved entirely within an AI answer. A high-consideration purchase may still lead users to product pages, reviews, demonstrations and company websites.

Google has also said that clicks originating from AI experiences can be higher quality and has continued testing designs intended to encourage website visits.

This is Google’s product position and should not be treated as an independent traffic conclusion.

Sources:

• Agarwal and Sen: AI Overviews field experiment
• Khosravi and Yoganarasimhan: AI Overviews and Wikipedia traffic
• Google: New controls for website owners

Are AI Citations the Same as Traditional Rankings?

No. AI citations and conventional search rankings overlap, but they are not the same system.

A 2026 measurement study issued 55,393 trending queries across 19 categories between March 13 and April 21.

AI Overviews appeared for:

• 13.7% of all queries
• 64.7% of question-form queries

Nearly 30% of the domains cited by AI Overviews did not appear in the accompanying first-page search results.

The researchers also broke responses into 98,020 individual claims and found that 11% were not supported by the cited pages.

This creates four separate visibility outcomes:

Brands should therefore avoid using ranking data as a substitute for AI visibility reporting.

Source: Measuring Google AI Overviews

How Did AI Visibility Measurement Improve?

Microsoft introduced one of the clearest first-party measurement changes in 2026.

In February, Bing Webmaster Tools launched AI Performance reporting in public preview.

It showed how frequently content was cited across Microsoft Copilot, Bing AI summaries and selected partner experiences, including which URLs were referenced.

On June 16, Microsoft added:

• Intents
• Topics
• Citation Share
• Competitor comparison

Citation Share indicates the proportion of citations attributed to a website among citations produced for the same grounding query.

Microsoft described the measure as observational rather than a direct ranking score.

A practical AI visibility dashboard should now include:

Because generated answers can vary between runs, priority prompts should be tested repeatedly rather than evaluated from one screenshot.

Sources:

• Microsoft: AI Performance in Bing Webmaster Tools
• Microsoft Bing Search updates

What Changed for Publishers and Content Owners?

June introduced meaningful regulatory and platform controls.

On June 3, the UK Competition and Markets Authority imposed a publisher conduct requirement on Google.

It required:

• Effective controls over the use of publisher content in generative AI
• Clearer explanations of how content is used
• Access to engagement metrics
• Reasonable steps toward accurate attribution

Google simultaneously began testing a Search Console control that allowed selected website owners to manage whether their sites could appear in and help ground AI Overviews, AI Mode and related generative features.

Google stated that websites opting out would not receive traffic or impressions from those AI experiences.

On June 17, the CMA imposed a separate fair-ranking requirement.

It covered organic results within generative AI features and required ranking criteria to be objective and non-discriminatory, alongside greater transparency around material changes.

These developments move publisher control, attribution and visibility reporting from optional product features toward regulatory expectations.

Sources:

• CMA: Google publisher conduct requirement
• CMA: Google fair-ranking requirement
• Google: New controls for website owners

What Should Brands Do Next?

Brands should continue investing in technical SEO, crawlability, structured information and topical authority.

AI visibility builds on those foundations rather than replacing them.

1. Map real customer prompts

Track questions across awareness, comparison, objection and purchase stages.

Do not limit monitoring to high-volume keywords. Include natural questions customers may ask an AI assistant before making a decision.

2. Measure multiple AI engines

Compare ChatGPT, Gemini, Google AI Search, Claude, Copilot and Perplexity separately.

A strong result on one engine does not guarantee visibility on another.

3. Publish original evidence

Use studies, benchmarks, expert commentary, unique examples and transparent methodologies that other sources can reference.

4. Strengthen entity consistency

Keep product descriptions, company details, leadership information and brand positioning consistent across the website and trusted external sources.

5. Audit citation accuracy

Check whether AI systems are linking to the right pages and representing the brand correctly.

Record inaccurate descriptions, outdated claims and incorrect product details.

6. Measure influence beyond clicks

Combine AI mentions, citations, referrals, branded search, direct traffic and assisted conversions.

AI visibility may influence a purchase before a measurable website session occurs.

CogNerd helps brands track mentions, citations, sentiment and competitor gaps across major AI platforms.

Explore the CogNerd AI marketing platform or start monitoring your AI visibility.

Methodology and Editorial Disclosure

This report reviewed official announcements published by Google, OpenAI, Microsoft and the UK Competition and Markets Authority between January 1 and July 30, 2026.

It also used traffic estimates from Similarweb and original or working academic research.

Company-reported usage statistics are labelled as such. Traffic figures from Similarweb are estimates. The studies cited may be revised or peer-reviewed after publication.

No undisclosed CogNerd client results were used in this article.

Practical recommendations represent CogNerd’s interpretation of the published evidence.

Primary Sources and Further Reading

• Google: Personal Intelligence in AI Mode
• Google: AI Mode and AI Overviews updates
• Google: New controls and insights for website owners
• OpenAI: Product discovery in ChatGPT
• OpenAI: New ways to buy ChatGPT ads
• Microsoft: AI Performance in Bing Webmaster Tools
• CMA: Google publisher conduct requirement
• Similarweb: AI Search Stats 2026
• Agarwal and Sen: AI Overviews field experiment
• Xu, Iqbal and Montgomery: Measuring Google AI Overviews

Thanks for Reading

Thank you for reading the 2026 AI Visibility Report.

AI search is changing quickly, but the strategic direction is already clear. Brands need to be discoverable not only in traditional search results, but also inside the answers, comparisons and recommendations that increasingly shape customer decisions.

CogNerd will continue monitoring platform developments, citation behaviour, publisher controls and the metrics marketers need to understand this new discovery environment.