Google AI Overviews Update History: What Changed and What It Means for Your Visibility
Google AI Overviews has evolved rapidly since its 2024 launch. This is a comprehensive, chronological record of every major update - what changed, what likely triggered it, and what it means for your site's AI search visibility. Bookmark this page; we update it as new changes roll out.
Understanding AI Overviews history isn't academic trivia. Each update shifted citation patterns. Sites that were visible before might have dropped. Others that were invisible suddenly started appearing. If you're not tracking these changes, you're flying blind in AI search.
Last updated: April 2026. We'll add new updates as Google rolls them out. Subscribe to our newsletter or set up continuous monitoring to get notified when changes affect your visibility.
2024: The Launch Year
Initial US Rollout
Google officially launched AI Overviews (previously Search Generative Experience / SGE) in the United States. AI-generated summaries began appearing at the top of search results for a subset of queries.
What changed for visibility:
- First-mover advantage established - early adopted sites that matched Google's initial criteria started getting citations
- Heavy weighting on traditional authority signals (backlinks, domain authority)
- Limited rollout - only ~15% of US queries showed AI Overviews
Key takeaway: The initial source selection favoured established sites with strong traditional SEO profiles. Newer sites struggled to appear regardless of content quality.
The "Eat Rocks and Glue" Correction
AI Overviews made headlines for suggesting people eat rocks for minerals and add glue to pizza. Google rapidly deployed corrections, tightening source quality requirements.
What changed for visibility:
- Stricter filtering of Reddit and forum content (previously weighted too heavily)
- Increased emphasis on EEAT signals for health and safety queries
- Satire and joke content more aggressively filtered
- Medical and health queries started requiring more authoritative sources
Key takeaway: Google demonstrated willingness to rapidly adjust citation criteria. The system was more volatile than traditional search - and would continue to be.
UK and EU Expansion
AI Overviews launched in the UK, Germany, France, Italy, Spain, and several other European markets.
What changed for visibility:
- Regional source preferences emerged - UK sites favoured for UK queries
- Local authority signals (ccTLD, regional backlinks) gained weight
- Currency and measurement localisation affected commerce queries
Key takeaway: AI Overviews wasn't globally uniform. Regional optimisation became important - what works for US visibility might not work for UK visibility.
On-Page Structure Recognition Improvement
Google updated how AI Overviews parsed and used schema.org structured data, particularly for FAQ, HowTo, and Product content.
What changed for visibility:
- Sites with comprehensive FAQPage schema saw citation increases
- HowTo schema became more important for instructional queries
- Product schema helped e-commerce sites appear in comparative queries
- Poorly implemented schema (missing required fields) started causing issues
Key takeaway: Structured data moved from "nice to have" to "significant ranking factor." Sites that had invested in proper schema implementation gained competitive advantage.
Citation Link Styling and Presentation Update
Google changed how source links appeared within AI Overviews, making them more prominent and easier to click.
What changed for visibility:
- Click-through rates to cited sources increased
- Sources with cleaner meta titles/descriptions saw higher engagement
- Multiple citations from the same domain became less common
Key takeaway: Being cited became more valuable as click-through improved. But competition for limited citation slots intensified.
2025: The Maturation Year
Reduction in "Show More" Expansion
Google reduced the number of sources shown in expanded AI Overviews, typically showing 3-4 primary sources instead of 5-7.
What changed for visibility:
- Competition for citations intensified - fewer slots available
- First-position sources gained disproportionate visibility
- Marginal sources (positions 5-7) largely lost their visibility
Key takeaway: Being "in the top 3" became critical. Fourth or fifth place often meant not being seen at all.
EEAT Integration Strengthened
Google integrated deeper EEAT (Experience, Expertise, Authoritativeness, Trustworthiness) signals into AI Overview source selection.
What changed for visibility:
- Author credentials became more important - named experts got preference
- Person schema for authors helped AI verify credentials
- Anonymous content or content without clear authorship dropped
- YMYL (Your Money Your Life) queries required demonstrably expert sources
Key takeaway: Content without clear authorship and credentials lost ground. The EEAT framework that affected traditional rankings now applied directly to AI citations.
On-Page Structure Requirements Tightened
Google raised the bar for structured data quality, penalising incomplete or incorrect schema implementation.
What changed for visibility:
- Schema with missing required fields started causing drops
- Inconsistent schema (mismatch between schema and visible content) penalised
- Article schema with proper datePublished and dateModified gained importance
- Organisation schema verification became stricter
Key takeaway: "Some schema" was no longer enough. Implementation quality mattered. Sites with comprehensive, accurate structured data pulled ahead.
llms.txt Recognition (Limited)
Google began experimentally recognising llms.txt files, joining Perplexity in supporting this emerging standard.
What changed for visibility:
- Sites with llms.txt saw modest citation improvements in some categories
- The file helped clarify site expertise areas
- Early adopters gained minor competitive advantage
Key takeaway: llms.txt moved from "Perplexity only" to broader relevance. Not yet critical, but increasingly worth implementing.
Mobile-First Source Selection
Google shifted AI Overviews source selection to prioritise mobile page experience, reflecting the dominance of mobile search.
What changed for visibility:
- Mobile Core Web Vitals became a factor in citation eligibility
- Sites with slow mobile load times saw visibility drops
- Mobile-unfriendly layouts hurt citation chances
- AMP pages didn't get special treatment - standard mobile experience mattered
Key takeaway: Desktop-optimised sites that neglected mobile fell behind. Mobile experience became a GEO factor, not just a ranking factor.
Brand Authority Verification Enhancement
Google improved how it verified brand authority, cross-referencing more external signals before citing sources.
What changed for visibility:
- Wikipedia/Wikidata presence became more important
- Consistent NAP (name, address, phone) across directories helped
- Third-party mentions and reviews factored into authority assessment
- New or unverifiable brands struggled to get citations
Key takeaway: Brand authority isn't just about your own site. External validation across the web matters. PR and third-party presence became GEO factors.
Query Intent Classification Refinement
Google improved how it classified query intent, showing different AI Overviews formats and sources for informational, commercial, and transactional queries.
What changed for visibility:
- Commercial queries started favouring product-focused content with pricing
- Informational queries emphasised educational, comprehensive content
- Local intent queries showed local business sources more prominently
- Same keyword could show different sources depending on inferred intent
Key takeaway: Content-query alignment became more nuanced. Optimising for the same keyword might require different approaches depending on intent.
2026: The Current State
Freshness Signal Amplification
Google increased weighting for content freshness, particularly for queries where timeliness matters.
What changed for visibility:
- Recently published/updated content got visibility boosts for time-sensitive queries
- Clear datePublished and dateModified in schema became more important
- Evergreen content on rapidly-changing topics needed more frequent updates
- "Last reviewed" signals started being parsed from content
Key takeaway: Static content ages faster in AI search. Regular updates and clear date signalling became competitive advantages.
Multimodal Content Recognition
AI Overviews began incorporating signals from images, videos, and other media when evaluating content comprehensiveness.
What changed for visibility:
- Content with relevant, well-labelled images saw improvements
- Video content (with transcripts/structured data) factored into citations
- Alt text and image schema became more important
- Text-only content on visual topics started underperforming
Key takeaway: "Content" now means more than text. Comprehensive coverage includes appropriate multimedia - properly labelled and structured.
Patterns and Implications
Looking across two years of updates, several patterns emerge:
1. Increasing Sophistication
Each update has made source selection more nuanced. Simple signals (backlinks, domain age) matter less; complex signals (EEAT, content structure, multimodal coverage) matter more.
2. Higher Quality Bar
The threshold for citation keeps rising. What qualified in May 2024 might not qualify today. Continuous improvement is required just to maintain position.
3. Convergence with GEO Best Practices
The updates consistently favour sites with strong GEO foundations: comprehensive schema, clear authorship, brand authority, mobile excellence, fresh content.
4. Regional Divergence
AI Overviews isn't globally uniform. Regional variations in source selection mean global brands need region-specific GEO strategies.
5. Competitive Intensification
Fewer citation slots, higher quality requirements, and increasing GEO awareness among competitors means the bar keeps rising.
What This Means for Monitoring
With updates this frequent and impactful, monitoring isn't optional. You need to:
- Track your score weekly - Catch drops before they compound
- Monitor competitors - See when they improve before they take your citations
- Correlate changes with updates - Understand whether drops are your issue or algorithm shifts
- Respond quickly - The faster you adapt to updates, the less visibility you lose
Stay ahead of AI Overviews changes
Weekly monitoring catches update impacts before they compound. Know when your visibility shifts - and why.
Start Monitoring →Looking Forward
Based on the trajectory, here's what we anticipate for AI Overviews in late 2026 and beyond:
- Deeper EEAT integration - More verification of claimed expertise
- llms.txt becoming standard - Moving from experimental to expected
- Real-time freshness - Near-instant indexing for breaking topics
- Cross-platform authority - Social signals, podcast appearances, conference talks factoring in
- Industry-specific variations - Different citation criteria for different verticals
The only certainty is continued change. AI Overviews will keep evolving, citation criteria will keep shifting, and the businesses that stay ahead will be those that monitor continuously and adapt quickly.
We'll update this page as new changes roll out. Bookmark it, or better yet, set up continuous monitoring so you know immediately when updates affect your visibility.
Continue reading: AI Visibility Monitoring
Sources & Further Reading
Check your AI visibility
Enter your URL at SearchScore for a free AI visibility score out of 100. See how ChatGPT, Perplexity and Google AI see your site - and exactly what to fix.