A new analysis of one million keywords has produced the clearest picture yet of how AI is transforming Google search. The findings are striking: nearly half of all tracked keywords are losing search volume, while a fifth are growing. The divide maps almost perfectly onto how AI Overviews interact with different types of search intent.
The Core Finding: A Search Volume Split
Researchers analyzing one million keywords across categories found:
- 40.7% of keywords show declining search volume - AI Overviews answer these queries directly, reducing the need for users to click through to websites
- 20.1% of keywords show growing volume - largely queries that AI tools have created demand for, or action-oriented searches where users still need to click
- 39.2% of keywords show stable volume - AI Overviews have limited presence for these query types
The split follows a clear pattern: informational queries ("what is X," "how does Y work") are the losers. Action and commercial queries ("buy X," "hire Y near me," "X pricing 2026") are the winners or stable performers.
Where AI Overviews Appear (and Where They Do Not)
AI Overviews now appear in an estimated 47-64% of all Google searches as of mid-2026, depending on the query category. Coverage varies dramatically:
High AIO presence (60%+ of queries):
- Health and medical information
- General how-to questions
- Historical and factual queries
- Definition and explanation searches
Low AIO presence (under 20% of queries):
- Local search ("near me," city + service)
- Commercial queries (product pages, pricing)
- Navigational queries (brand names, website lookups)
- Time-sensitive news and events
Medium AIO presence (20-60%):
- Financial and legal information
- Educational content
- Product research and comparison
The CTR Impact Is Real and Measurable
Position 1 organic CTR has dropped from approximately 27% (pre-AIO baseline) to approximately 11% in categories with high AIO coverage. This is not universal - in low-AIO categories, position 1 CTR remains near historical norms.
The practical implication: a site ranking position 1 for a high-AIO informational query in 2026 receives roughly 40% of the clicks it would have received for the same position in 2024.
This is not a Google quality issue. AI Overviews are answering questions accurately. Users are simply not clicking because they received the answer without needing to.
Which Content Types Are Surviving
The keyword analysis reveals which content formats retain click-through even when AI Overviews are present:
Surviving formats:
- Original research with data and statistics that AI cannot summarize without linking to the source
- Step-by-step implementation guides where the user needs to follow along
- Time-sensitive analysis ("what just happened with X," "how does [recent event] affect Y")
- Comparison and review content where trust in the publisher matters
- Commercial and transactional pages - AI Overviews rarely appear on shopping intent queries
Declining formats:
- Evergreen "what is X" and "how X works" articles
- Definition pages and glossary entries
- General overview articles without original data
- Lists without unique research or curation
What This Means for Content Strategy in 2026
The data supports a content strategy pivot that many SEO practitioners have already observed but struggled to quantify:
- Stop writing for definitions. "What is [topic]" queries now produce AIO responses in the majority of cases. Unless your content adds original research, first-person experience, or proprietary data that AI cannot access, ranking for these terms no longer generates meaningful traffic.
- Write for decisions. Content that helps users make a decision - what to buy, which vendor to hire, whether to take an action - retains click-through because AI Overviews are reluctant to make explicit recommendations. Decision-support content is defensible.
- Data is a moat. Original statistics, survey results, and research data attract both links and AI citation. When AI cites your data, it often drives clicks to verify the source. This is the most durable form of informational content in the AIO era.
- Action-oriented content survives. "How to do X step by step" retains value better than "what is X" because users following a process need to return to the source repeatedly. Instructional content with screenshots, code, or detailed procedure is harder for AI to fully substitute.
The Growing Keywords: What AI Is Creating
The 20.1% of keywords showing volume growth fall into recognizable categories:
- AI tool and workflow queries: "How to use ChatGPT for [task]," "Claude vs GPT for [use case]," "best AI tools for [role]" - a new category of queries that did not exist two years ago
- Verification queries: Users are increasingly searching to verify or fact-check AI outputs - "is [AI claim] true," "[AI stated fact] verify"
- Prompt and workflow searches: "Prompt for [task]," "AI workflow for [business process]"
- AI-in-category queries: "AI [existing category]" variants are growing across virtually every vertical
For content publishers, these growth categories represent real opportunities - particularly the intersection of AI tools and specific professional domains.
Implications for SEO Strategy
The one million keyword analysis confirms a fundamental shift in how organic search generates value. Traffic volume from informational content will continue to decline as AIO coverage expands. The sites that maintain and grow organic traffic in 2026 and beyond are those that:
- Produce original data and research (cannot be replicated by AI without attribution)
- Serve action and commercial intent (where AIO is weakest)
- Build topic authority deep enough to be cited by AI rather than replaced by it
- Focus on engagement depth (time-on-site, return visits) over raw click volume
The search optimization game has not ended - it has shifted from capturing informational queries to owning decision-stage and action-stage traffic.
