Two AI Search Features, One Google

Google is running two distinct AI search systems simultaneously, and most SEO teams are treating them as one thing. AI Overviews are the automatic AI-generated summaries that appear at the top of roughly 30% of search results. AI Mode is a separate, opt-in conversational interface that users actively choose to enter. Same search box, same company, radically different technology under the hood. The content that gets cited in one is largely ignored by the other.

This distinction matters because each feature uses different retrieval mechanisms, favors different content formats, and serves different user intents. AI Overviews act as a concise answer layer on top of the traditional SERP. They are one-shot: the user asks a question, gets a summary, and either clicks a citation or refines their search. AI Mode operates as a multi-turn conversation powered by Gemini, with the ability to ask follow-ups, dig deeper, and even access personal data through Google's Personal Intelligence integration with Gmail, Photos, and YouTube history.

The practical consequence is that a single "AI search optimization" strategy is no longer sufficient. You need two separate playbooks: one for winning citations in the concise, authority-driven AI Overviews, and another for earning visibility in the deep, conversational AI Mode. Our AI search optimization guide covers the broader landscape, but this post focuses on why these two features diverge so sharply and how to build a strategy for each.

The 13.7% Overlap: What the Ahrefs Data Reveals

Ahrefs ran the most comprehensive study to date, analyzing 730,000 AI-generated responses across both AI Mode and AI Overviews. The headline finding: only 13.7% of cited domains overlap between the two features. That is not a rounding error. It means 86.3% of the sources cited in one feature are not cited in the other. A site that dominates AI Overviews citations may be completely invisible in AI Mode, and vice versa.

This low overlap is driven by fundamental differences in how each feature retrieves and ranks information. AI Overviews lean heavily on high-authority domains that already rank well in traditional organic search. They favor concise, well-structured content with clear factual claims. AI Mode, by contrast, casts a wider net. It pulls from more diverse sources, values depth over brevity, and gives weight to content that covers a topic from multiple angles. The result is two largely separate citation ecosystems operating within the same search engine.

The 13.7% overlap also tells us something about the user intent each feature serves. AI Overviews answer direct questions quickly. AI Mode supports exploratory research where the user does not yet know exactly what they need. This intent divergence means Google's retrieval system is selecting different types of sources for different types of information needs, even when the initial query looks similar. If you want to understand how different AI platforms handle citations differently, our AI citation optimization guide breaks it down across ChatGPT, Perplexity, and Google.

How AI Overviews Work (and Who They Cite)

AI Overviews are triggered automatically when Google determines a query would benefit from a synthesized answer. They currently appear for approximately 30% of all searches, skewing heavily toward informational queries but increasingly appearing for commercial and "best of" queries as well. The user does not opt in. If Google decides your query warrants an AI Overview, it appears above the organic results whether you want it or not.

The retrieval mechanism behind AI Overviews is tightly coupled with Google's organic search index. Pages that already rank in the top 10 for a given query are disproportionately represented in AI Overview citations. Google's system pulls snippets from these high-ranking, high-authority pages and synthesizes them into a cohesive answer. The citations appear as small source links, typically 3 to 5 per Overview. Our AI Overviews optimization guide goes deep on the mechanics, but the key insight here is the authority bias. If you do not already have strong organic rankings and high domain authority, your odds of appearing in AI Overviews citations are slim.

Content format matters significantly. AI Overviews favor pages with clear, direct answers near the top of the content, followed by supporting detail. Think of the inverted pyramid structure from journalism: lead with the answer, then provide context. Pages that bury their key insights under lengthy introductions or require readers to scroll through thousands of words before reaching the point get passed over. Google's AI wants to extract a clean, authoritative snippet, and it gravitates toward content that makes that extraction easy. Structured data for AI search is a critical enabler here, because schema markup helps Google's system understand exactly what your content claims and where the key facts live.

The one-shot nature of AI Overviews also means that conciseness wins. The AI Overview itself is typically 100 to 250 words. Google needs to pull tight, factual, well-sourced claims, not sprawling essays. Sites that structure their content with clear headings, direct answer paragraphs, and supporting data tables consistently outperform those that rely on long-form narrative alone. This is the opposite of what works in AI Mode, which is why a single strategy cannot serve both features.

How AI Mode Works (and Who It Cites)

AI Mode is a fundamentally different product. Users opt into it by clicking the AI Mode toggle in Google Search, entering a conversational interface powered by Gemini. Unlike AI Overviews, which are a one-shot answer, AI Mode supports multi-turn conversations with follow-up questions, clarifications, and progressive deepening of a topic. The user might start with "What are the best CRM tools for mid-market SaaS?" and follow up with "How does HubSpot compare to Salesforce on pricing?" and then "What about data migration from one to the other?" AI Mode handles the entire thread, building context across the conversation.

The retrieval system for AI Mode is more expansive than AI Overviews. Where AI Overviews lean on top-ranking, high-authority pages, AI Mode cites a wider range of sources including niche blogs, forum discussions, specialized publications, and long-tail content that rarely appears on page one of traditional search results. The Ahrefs data confirms this: AI Mode cites more diverse domains, pulling from the mid-tier and long-tail of the web rather than concentrating citations among a handful of dominant publishers. Our Google AI Mode SEO guide covers the optimization mechanics in detail.

AI Mode's Deep Search feature is particularly notable. When activated, it performs a significantly more thorough retrieval pass, digging into subtopics and pulling supporting evidence from sources that a standard search would never surface. For complex, multi-faceted queries, Deep Search can reference 20 or more sources in a single response. This creates opportunities for content that goes deep on specific subtopics, even if it does not have the domain authority to rank on page one organically.

The Personal Intelligence layer adds another variable entirely. AI Mode can access a user's Gmail, Google Photos, and YouTube history to personalize responses. This means two users asking the same query can see different cited sources based on their personal context, past searches, and content preferences. Personalization makes AI Mode citation tracking harder and reduces the reliability of rank-tracking tools for this feature. It also means that brand familiarity, via email newsletters, YouTube content, and past site visits, can influence whether AI Mode surfaces your content for a given user.

Response Length and Depth Differences

AI Mode responses are 4x longer than AI Overviews. That is not a stylistic choice; it reflects a fundamentally different information architecture. AI Overviews are designed to answer a query in the fewest words possible, typically 100 to 250 words, with 3 to 5 inline citations. They are a summary layer. AI Mode responses regularly exceed 800 to 1,000 words per turn, with 10 to 20+ citations, and they grow longer when the user asks follow-up questions. A multi-turn AI Mode session can produce 3,000 or more words of synthesized content.

This length difference has direct implications for what content gets cited. AI Overviews need clean, extractable snippets. A single well-structured paragraph with a clear factual claim is often enough to earn a citation. AI Mode needs depth. It is synthesizing nuanced explanations, comparing alternatives, and presenting evidence from multiple angles. Content that goes deep on methodology, includes case studies, presents original data, and covers edge cases is what AI Mode draws from. Thin content with surface-level coverage gets filtered out entirely.

The depth difference also means that AI Mode values topical coverage across a site, not just on a single page. If your site has 15 well-researched articles covering different aspects of a topic, AI Mode may cite 5 or 6 of them across a multi-turn conversation. This is the hub-and-spoke content model that our content strategy service builds for clients: a central pillar page supported by detailed subtopic pages, each going deep enough to serve as a standalone citation source. AI Overviews, by contrast, almost always cite the single most authoritative page on the topic, which is usually the pillar.

Different content formats also perform differently across the two features. Tables, comparison matrices, and data visualizations are particularly effective for AI Overviews because Google can extract structured data points cleanly. AI Mode, on the other hand, draws more heavily from explanatory prose, expert analysis, and narrative-style content that provides context and reasoning behind the facts. The ideal dual-optimization approach creates content that layers both formats: clear data structures for AI Overviews extraction plus in-depth analysis for AI Mode synthesis.

The Click-Through Rate Gap: 43% vs 93%

The zero-click rates tell different stories about different futures. AI Overviews produce a 43% zero-click rate, meaning 57% of users who see an AI Overview still click through to an external website. That is a significant traffic reduction compared to traditional organic results, but more than half of users still find reason to visit a source page. The citations within AI Overviews function somewhat like featured snippets: they reduce click-through but do not eliminate it. For our zero-click searches guide, AI Overviews represent the moderate scenario.

AI Mode is the extreme scenario. A 93% zero-click rate means fewer than 1 in 10 AI Mode searches produce an external click. The reason is structural: AI Mode's conversational follow-up system is designed to keep users inside the Google interface. When the AI answers your question and immediately suggests related follow-up questions, the natural behavior is to continue the conversation, not click out to a website. Each follow-up generates more AI content, which generates more follow-up suggestions. It is a retention loop that makes external clicks a deliberate detour rather than a default action.

The traffic implications are asymmetric depending on which feature you optimize for. Winning an AI Overviews citation for a high-volume query still delivers meaningful referral traffic. The 57% click-through rate on a query with 50,000 monthly searches means real visitors. Winning an AI Mode citation for the same query delivers almost no direct traffic, because only 7% of users click through and AI Mode adoption is still growing. But the brand impression from that AI Mode citation reaches every user in the conversation. Run your key pages through our SERP Preview Tool to see how they currently appear across both features.

This gap means your KPIs should differ by feature. For AI Overviews, referral traffic and click-through rate remain meaningful metrics. For AI Mode, brand mention frequency, citation share, and branded search lift are the metrics that matter. Trying to evaluate AI Mode performance by traffic alone will make it look like a failure, even when it is generating substantial brand exposure. Understanding this distinction is the foundation of any honest AI search strategy.

Brand Visibility: 90% vs 43%

Brands appear in 90% of AI Mode responses but only 43% of AI Overviews. This is one of the most underappreciated findings in the Ahrefs study, and it flips the conventional assumption about which feature is more valuable for brand building. AI Mode's longer, more detailed responses naturally reference brand names as it discusses products, services, comparisons, and recommendations. When a user asks AI Mode to compare CRM platforms, the response will name HubSpot, Salesforce, Pipedrive, and others by name, often multiple times. AI Overviews, with their compressed format, frequently summarize information without attributing it to specific brands.

This means AI Mode is a significantly more powerful brand visibility channel than AI Overviews, despite its much lower click-through rate. Every AI Mode conversation that mentions your brand is a micro-branding event. The user sees your name associated with expertise, recommendations, or positive comparisons. Over time, this accumulates into brand recognition that drives branded searches, direct visits, and higher conversion rates through other channels. This is the same dynamic we outline in our author entity E-E-A-T guide, but operating at the brand level rather than the individual level.

The 43% brand mention rate in AI Overviews still matters, particularly for commercial queries where Google includes brand names in product comparisons and recommendation summaries. But the disparity suggests that AI Mode is where brand wars will be fought in 2026 and beyond. The brands that figure out how to consistently appear in AI Mode's conversational responses will build compounding awareness advantages that show up as branded search volume increases months later.

The Search Everywhere Optimization playbook we published covers how to track brand mentions across AI platforms, but the specific AI Mode strategy is about ensuring your brand is mentioned in the right context. Being cited as a budget alternative is different from being cited as the industry leader. The content you create, the reviews you earn, and the third-party mentions you generate all shape how AI Mode characterizes your brand in its responses.

Optimizing for AI Overviews: The Concise Authority Strategy

AI Overviews citation is won through authority and structure. The first prerequisite is strong organic rankings: if you are not in the top 10 for a target query, your chances of being cited in the AI Overview are close to zero. This is not the feature where underdogs win. Domain authority, topical authority, and existing SERP presence are the foundation. Start by auditing which of your target queries trigger AI Overviews and whether you currently rank in the top 10 for those queries. Our SEO Score Calculator helps identify these gaps.

Once you have the authority base, the content structure is what determines whether Google's AI extracts your content as a citation source. Lead every page with a direct answer to the primary query in the first 100 words. Use clear H2 and H3 heading structures that match common question patterns. Include specific numbers, dates, and data points that the AI can reference as factual claims. Implement Article, FAQ, and HowTo schema markup so Google's retrieval system can parse your content programmatically. The Schema Markup Generator in our tools suite handles the technical implementation.

E-E-A-T signals carry outsized weight in AI Overviews. Google's AI is cautious about what it presents as factual, so it preferentially cites sources with demonstrated expertise: author bios with credentials, editorial policies, cited sources within the content, and established publication history on the topic. This is where a comprehensive SEO audit pays for itself. If your E-E-A-T signals are weak, no amount of content optimization will overcome Google's trust filter for AI Overviews citations.

The format that consistently wins AI Overviews citations is what I call the "answer-then-expand" pattern. Start with a definitive 2-3 sentence answer. Follow with a short supporting paragraph that adds nuance. Include a data table or comparison if the query involves evaluating options. Then go deeper in subsequent sections. This layered approach gives Google's AI a clean snippet to extract for the Overview while also providing the depth signals that maintain your organic rankings. Pages that use this pattern see 2-3x higher AI Overview citation rates than those that bury their answers in the middle of long-form content.

Optimizing for AI Mode: The Deep Expertise Strategy

AI Mode optimization rewards the opposite of what works for AI Overviews. Where AI Overviews want concise authority, AI Mode wants deep expertise. The content that earns AI Mode citations is comprehensive, multi-dimensional, and rich with original insight. Think 3,000+ word guides that cover methodology, edge cases, practical examples, and expert commentary. Think original research with proprietary data that AI Mode cannot find anywhere else. Think content that answers not just the primary question but the five follow-up questions the user is likely to ask next.

The domain authority bar is lower for AI Mode, which is good news for smaller publishers and niche sites. Because AI Mode's retrieval system casts a wider net, you do not need to be a DR 80+ domain to earn citations. What you need is genuine depth on your topic. A specialized blog with 50 well-researched articles about B2B marketing attribution will get cited in AI Mode conversations about marketing analytics, even if it never appears on page one of traditional Google results. This is where topical authority at the site level, not just the page level, becomes your competitive advantage.

AI Mode's Deep Search feature is your opportunity multiplier. When users activate Deep Search, Google's retrieval digs significantly deeper into its index, surfacing content that standard search would never rank. Creating content specifically designed for Deep Search means going where mainstream content does not: detailed methodology sections, contrarian analysis with supporting evidence, step-by-step implementations with code examples or templates, and nuanced comparisons that cover the trade-offs most content skips. Use the AI Content Optimizer to evaluate whether your content has the depth signals AI Mode looks for.

The multi-turn conversation model also means you should structure content to anticipate conversational follow-ups. If your pillar page covers "enterprise SEO strategy," you should have supporting pages on "enterprise SEO team structure," "enterprise SEO tool stack," "enterprise SEO budget allocation," and "enterprise SEO reporting frameworks." When a user asks AI Mode about enterprise SEO and then follows up about team structure, AI Mode will look for a dedicated resource on that subtopic. If your site has one, you earn the citation. If it does not, a competitor's site will. This is the hub-and-spoke architecture that our AIO optimization service builds for every client engagement.

Building a Dual Optimization Workflow

Running two separate optimization strategies does not mean doubling your workload. It means structuring your content creation process to serve both features from a single production pipeline. The most efficient approach starts with a content architecture audit: map every target query to whether it triggers an AI Overview, is likely to be asked in AI Mode, or both. Queries where both features are relevant get the dual-format treatment. Queries that only trigger AI Overviews get the concise authority format. Long-tail, complex queries that are natural AI Mode conversations get the deep expertise format.

For dual-format pages, the structure is specific. Lead with a 150-word direct answer section optimized for AI Overviews extraction, including schema-marked FAQ entries near the top. Then expand into comprehensive, expert-depth sections that serve AI Mode's preference for thorough coverage. Include comparison tables and data summaries (AI Overviews loves extractable structured data) alongside the detailed analysis paragraphs that AI Mode synthesizes from. This layered approach lets one page serve both retrieval systems without compromising either. Test your pages through our AIO Readiness Checker to verify they meet both sets of requirements.

Measurement must also be split. Track AI Overviews performance through citation monitoring, click-through rate from AI Overview results, and organic traffic trends for queries where AI Overviews appear. Track AI Mode performance through brand mention monitoring across conversational search sessions, branded search volume growth, and citation diversity metrics. Google Search Console is starting to separate AI Mode impressions from standard search impressions, which gives you the data infrastructure to build split reporting. Our optimization consultation includes setting up this dual measurement framework from day one.

The workflow cadence matters too. AI Overviews citations respond relatively quickly to content changes because they are tied to existing organic rankings. Update your page, improve your featured snippet position, and your AI Overviews citation can shift within weeks. AI Mode citations have a longer feedback loop because they depend on topical authority signals that build over time: the breadth of content on your site, the diversity of subtopics covered, third-party references, and accumulated expertise signals. Plan AI Overviews optimization as a quarterly sprint and AI Mode optimization as a 6-month sustained effort. Both are essential. Neither is optional. The brands that build both pipelines now will own the AI search landscape that every competitor scrambles to figure out in 12 months.

Frequently Asked Questions