How AI Is Changing SEO: Google AI Overviews, AI Mode and Search Strategy
- Observational Keyword Modeling (Ahrefs): In April 2025, Ahrefs analyzed 300,000 keywords and estimated that the presence of an AI Overview correlated with a 34.5% lower average click-through rate (CTR) for the top-ranking page compared to similar queries without one. In a February 2026 follow-up analysis, Ahrefs’ model estimated that the position-one CTR difference correlated with AI Overviews had widened to 58%. Because these figures are comparison-based estimates across specific keyword samples, they do not mean every website or number-one ranking will lose 58% of its clicks—branded, transactional, and navigational queries behave very differently.
- Behavioral Panel Data (Pew Research Center): A July 2025 Pew Research Center study analyzed 68,879 Google searches from U.S. participants during March 2025. Participants clicked a traditional search link in 8% of visits to search pages with an AI summary, compared to 15% of visits to search pages without one. Users clicked a link inside the AI summary itself in 1% of visits, and 26% ended their browsing session entirely after viewing a page with an AI summary (compared to 16% on standard results pages).
AI has changed how search engines retrieve and synthesize information, as well as how marketing teams research, write, and measure pages. Adapting to this shift starts with solid SEO services and extends into structured editorial practices often grouped under Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO).
Key Takeaways
- AI summary frequency depends on query length and intent: In Pew Research’s behavioral study, AI summaries appeared on 18% of all tracked searches, rising to 53% for queries of 10 or more words and 60% for questions starting with words like what, why, or how. Similarly, Ahrefs’ 300,000-keyword study found that 99.2% of queries triggering an AI Overview had informational intent.
- Google relies on multiple distinct AI systems: RankBrain and BERT operate as core ranking systems, MUM is applied to specific search features rather than general page ranking, and AI Overviews and AI Mode use generative models and techniques that can vary by query.
- Google evaluates content value and purpose, not just the tool used: Using AI to assist with research or drafting is permitted, whereas mass-producing unoriginal pages that offer little value to users violates Google’s Scaled Content Abuse policy.
- Technical and editorial clarity supports eligibility, not guaranteed citations: Crawlability, clear answers, and accurate structured data help search systems interpret your pages and support eligible search features, but no formatting formula or schema markup guarantees inclusion in an AI Overview.
How Google’s AI Systems Actually Work (2015 to Present)
Older SEO guides often treat Google’s AI as a single algorithm—or mistakenly describe RankBrain as a standalone search engine. According to Google’s Search Ranking Systems documentation, Search relies on separate machine-learning systems that perform distinct functions:
- RankBrain (2015): A core ranking system that helps Google relate words to broader concepts. It allows Search to surface relevant pages even when they do not contain the exact words used in a query.
- BERT (2019): Short for Bidirectional Encoder Representations from Transformers, BERT is a ranking system that processes words in relation to all other words in a sentence rather than one by one in order. This helps Google interpret how context and prepositions alter query meaning.
- MUM (2021): Short for Multitask Unified Model, MUM is trained across 75 languages and multiple modalities. Unlike RankBrain and BERT, Google notes that MUM is used only for specific search applications—such as improving featured snippet callouts—rather than acting as a general web ranking system.
- AI Overviews and AI Mode: Google uses generative models, including custom Gemini models, to power AI Overviews and AI Mode. While both synthesize information from indexed web pages using techniques like “query fan-out” (issuing multiple related sub-queries behind the scenes), Google’s developer documentation explicitly notes that AI Mode and AI Overviews may use different models and techniques, meaning the responses and cited links they display will vary. AI Overviews trigger when Google’s systems determine a summary adds value beyond classic search results, whereas AI Mode is designed for deeper, multi-turn conversational follow-ups.
SEO vs. AEO vs. GEO: Industry Frameworks Explained
In its official guide to optimizing for generative AI features in Google Search, Google explains that established SEO best practices remain foundational to visibility in AI-powered Search experiences, and that optimizing for generative AI search is still fundamentally SEO.
In day-to-day agency work, however, marketers often use Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) as practical industry frameworks—not official Google ranking systems—to organize how teams work across classic results, direct answer boxes, and third-party AI assistants.
| Dimension | Traditional SEO | Answer Engine Optimization (AEO) (Industry Framework) | Generative Engine Optimization (GEO) (Industry Framework) |
|---|---|---|---|
| Primary Goal | Rank a URL in classic organic search results | Improve eligibility for direct answer boxes, Featured Snippets, and voice responses | Earn brand mentions and source citations across generative AI platforms |
| Target Surfaces | Google Web Search, Bing | Google AI Overviews, Featured Snippets, People Also Ask | ChatGPT Search, Perplexity, Google AI Mode, Claude, Copilot |
| Common Focus Areas | Crawlability, internal linking, backlinks, page experience | Clear question-and-answer structure, valid schema, concise definitions | Consistent brand/entity details, original data, third-party citations |
| Typical Metrics | Organic clicks, rankings, and site conversions | Generative AI impressions and snippet visibility | Brand mention share and referral traffic from AI tools |
None of the focus areas above act as a guaranteed citation trigger. Instead, combining foundational technical SEO with structured AI SEO practices—and extending into Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) where your buyers research in conversational tools—reduces ambiguity when systems evaluate your pages.
Where AI Genuinely Helps SEO Operations (And Where to Be Careful)
Using AI tools inside an SEO workflow can cut days of manual analysis down to hours, provided human editors remain accountable for accuracy and usefulness.
1. Keyword Clustering and Programmatic Pages
Grouping thousands of search queries by intent, auditing indexation bloat, and spotting keyword cannibalization across a site are ideal tasks for machine-learning tools. Similarly, when a business maintains a valuable proprietary dataset—such as verified product specifications, live inventory, or location-specific service data—programmatic SEO can structure that information into useful landing pages. The deciding factor is user value: if programmatic templates spin out near-duplicate pages without distinct data, they risk violating Google’s Doorway or spam policies.
2. Content Creation and Google’s Scaled Content Abuse Policy
Google Search Central’s guidance on AI-generated content explains that Google focuses on the quality and helpfulness of content rather than how it was produced. Using AI to assist with outlining, summarizing data, or drafting is not inherently a violation.
What violates Google’s web search spam policies is Scaled Content Abuse: generating many pages primarily to manipulate search rankings rather than help users, typically producing unoriginal or low-value content regardless of whether it was created by AI, humans, or a combination of both. AI works best as a research and structuring assistant while human subject-matter experts contribute original pricing math, testing notes, and real trade-offs.
3. Link Analysis vs. Automated Outreach
Mass-automated outreach often produces generic messages and can damage a site’s reputation when used carelessly. AI is more useful for researching relevant websites, identifying broken references, and organizing original data that may support genuine editorial links.
4. Multimodal Search Readiness
Through Google Lens and Circle to Search, users frequently search by snapping a photo or circling an object on their screen and asking a follow-up question. Supporting your text with original, high-resolution images and videos, keeping Product structured data matched to visible page text, and maintaining accurate Google Merchant Center and Business Profile details helps your pages surface in visual and multimodal queries.
Editorial & Technical Practices for AI Search Eligibility
Google’s documentation emphasizes that even when a page meets all technical requirements and follows every best practice, indexing and serving are never guaranteed. However, following consistent editorial and technical standards makes your content easier for both human readers and search systems to parse:
- Ensure that Google can crawl and index your priority pages: Check
robots.txt, CDN or hosting restrictions, indexing directives, canonical URLs, and whether important content is available as accessible text. Use Search Console to inspect URLs and diagnose technical issues. Pages must be indexed and eligible to show a snippet in Search. Google’s Search generative AI control is available in Search Console. Inclusion is the default, but check that your property has not been set to exclude its content from AI Overviews, AI Mode, or generative AI features in Discover. - Lead sections with clear, direct explanations: Answering the core question clearly near the top of a section—before diving into secondary caveats—helps both scanning readers and passage-retrieval systems identify your main point.
- Use semantic HTML, lists, and tables where appropriate: While Google can process imperfect HTML, organizing multi-variable comparisons, pricing tiers, and specifications into clean HTML tables and lists improves readability across devices and assistive technologies.
- Use structured data when it accurately describes visible content: Implement Schema.org markup (such as
Organization,ProfessionalService,Article, orProduct) when it reflects the visible page text and supports an active, eligible search feature, and validate it with Google’s Rich Results Test. (Note that Google retiredFAQPagerich results in May 2026.) Structured data does not guarantee rankings or AI citations, and there is no special Schema.org type required for AI Overviews or AI Mode. See our small business SEO checklist for a foundational technical review. - Focus on non-commodity information and page experience: Google notes that clicks from results pages with AI Overviews have been associated with users spending more time on a site. A clear mobile layout, useful original information, and a fast page experience help visitors find the information they need. Measure engagement and conversions rather than assuming that AI-referred visitors automatically have greater purchase intent.
How to Measure AI Search Visibility and Leads
Before editing existing pages or launching a new content sprint, establish a clear measurement workflow so you can separate real progress from guesswork:
1. Check the Generative AI Performance Report in Search Console: In June 2026, Google launched dedicated Search Generative AI performance reports in Search Console for Search and Discover. Use this report to monitor:
- Impressions: How often links to your site were shown in generative AI features (including AI Overviews and AI Mode).
- Pages: Which specific canonical URLs are surfacing in AI features.
- Countries and devices: Review country-level visibility in the relevant report. Device-level breakdowns are available in the Search generative AI performance report, not the Discover report.
- Dates: Impression trends across daily, weekly, and monthly views.
2. Record a query and traffic baseline in classic Search Console: Because overall clicks and impressions from AI features are included inside the standard Web search type in the main Performance report, log your baseline clicks, impressions, and average positions for your priority queries before making on-page changes.
3. Audit brand citations across third-party AI platforms: Because Google Search Console only covers Google surfaces, maintain a recurring test sheet of 20 to 30 buyer questions and record whether your brand or URLs are cited in ChatGPT Search, Perplexity, Claude, and Microsoft Copilot.
4. Track engagement and conversions in Analytics and CRM: Evaluate page performance by tracking engaged sessions, time on page, form submissions, and qualified leads in Google Analytics 4 and your CRM alongside raw traffic numbers.
Frequently Asked Questions
Does Google penalize AI-generated SEO content?
No. Google does not penalize content simply because AI tools were used to help create it. Google targets Scaled Content Abuse—producing large volumes of unoriginal, low-value pages primarily to manipulate rankings rather than help users, whether created by AI, humans, or a mix of both.
How can I track Google AI Overview and AI Mode performance in Search Console?
You can use the Generative AI performance report in Google Search Console to view impressions, pages, countries, devices (available in the Search report, not the Discover report), and date trends for URLs appearing in AI Overviews and AI Mode. Overall clicks and impressions from these features are also included within the standard Web search type in the main Search Console Performance report, which you can pair with Google Analytics to measure on-site engagement and conversions.
Do I need special AI text files or proprietary schema to appear in Google AI Overviews?
No. Google’s documentation explicitly states that you do not need to create new machine-readable files, custom AI text files, or special Schema.org types to appear in AI Overviews or AI Mode. Standard Search indexing and snippet eligibility, valid structured data where appropriate for supported search features, and helpful, non-commodity content govern eligibility.
