LATEST UPDATES

A revised University of Washington study estimates that Wikipedia lost roughly 5% of Google referral traffic after AI Overviews. The paper is not yet peer-reviewed and Google disputes whether the analysis isolates AI Overviews, so the result is a directional signal rather than a universal benchmark.

BrightEdge reports that AI Mode cites Google-owned surfaces far more often than AI Overviews. The vendor has not published raw citation counts or the full sample, so treat the figure as an attributed market signal rather than Google-confirmed fact.

Profound found that persona details can materially change brand mentions and citations across ChatGPT, Claude and Gemini. The study supports testing prompts by audience segment, while its commercial source means teams should validate the pattern in their own data.

Merchant Center now exposes AI-search intent, search terms and missing product attributes in AI Performance Insights. Ecommerce teams can use the report to connect product data quality with visibility in AI Mode and AI Overviews.

Two third-party datasets indicate that AI Overviews now dominate People Also Ask answers. The sample is large, but Google has not confirmed the change and the methodology is not fully disclosed.

Google’s limited exact- and phrase-match ad test suggests AI Mode is carrying stronger commercial intent. The ad launch is Google-confirmed; the implications for organic competition remain practitioner analysis.

A practical test challenges the idea that AI systems inherently treat schema markup as privileged retrieval input. Because this is one self-published experiment, it should guide further testing rather than settle the question.

The method uses fan-out queries to identify external entities and sources that repeatedly shape AI recommendations. It is a practical discovery workflow, though performance figures come from the vendor’s own data.

Common Crawl data shows that many llms.txt files are implemented as if they were robots.txt, limiting their usefulness. The analysis is grounded in open crawl data and highlights concrete implementation mistakes.

The study suggests ChatGPT’s shortlist can influence product choice before a click, with comparison tables, pricing and “best for” labels shaping decisions. The sample is small and commercial, so the findings show correlation rather than causation.

Two studies show that search intent increasingly spans platforms, formats and tasks rather than a single query-result journey. The analysis combines Google usage data with external SERP and eye-tracking research and should be read with each study’s methodological limits.

The interview connects crawlability, rendering, internal linking, knowledge graphs and clear page intent to AI visibility. It is expert guidance rather than confirmed platform-factor documentation, but offers a useful technical audit framework.