The evolving AI-era search landscape shows Google’s AI-mode citations and the March core update data signaling a shift toward origin content over aggregators, making agent-ready site improvements essential for future traffic.

Google fixed a logging error that affected impressions from mid-2025 to early 2026; however, historical data will not be corrected, so impressions, CTR, and position from that window should be treated as unreliable for year-over-year comparisons.

Global rollout of Preferred Sources lets users mark sites to appear more in Top Stories, with higher click-through for marked sources, acting as a user-level trust signal rather than a sitewide ranking boost.

AI Mode citations surged to about 40 million by early May 2026, underscoring YouTube becoming a major source of citations in AI responses rather than just a platform for SEO.

Google added five AI Mode and AI Overview updates to improve citation presentation — inline links, desktop hover previews, end-of-response topic angles, and subscription-highlighted sources from followed outlets; creator names/handles now appear when social content is cited.

March core update findings show aggregators losing ground while brand and first-party sites gain; YouTube suffered the largest one-domain drop, with other declines seen on Reddit, Instagram, and X, indicating a shift toward content ownership over hosting platform.

Google released an agent-friendly website guide emphasizing stable layouts, semantic HTML, and accessibility, noting agents read the DOM and accessibility tree data before rendering screenshots.

DesignRush’s weekly SEO roundup ties together the March core update, data-logging issues, and AI-mode/citation changes to explain impacts on organic visibility and AI discovery.

Analysts identify agent-friendly design as a critical pre-I/O priority, with three immediate actions: audit the affected Search Console data window, assess March core update patterns for aggregator content, and implement agent-friendly optimizations like stable layouts, semantic HTML, and accessibility.