When Google launched its Search Generative Experience in the US and then expanded AI Overviews globally, the SEO industry went through a predictable cycle: alarm, overreaction, and eventual recalibration. We are now firmly in the recalibration phase — which means it is possible to assess the actual situation rather than reacting to speculation.
The honest picture is more nuanced than either the doom predictions or the dismissive reassurances suggested. AI Overviews do take organic clicks from some query types. They also create new opportunities for sites that position their content to be cited by AI. And the fundamentals of what makes content rank and earn trust have not changed as dramatically as many feared.
Here is what you actually need to know and do.
What Has Actually Changed With AI Overviews
Google's AI Overviews (formerly Search Generative Experience or SGE) use a large language model to generate a synthesised answer to search queries, displayed prominently above the traditional organic results. The AI draws on indexed web content to construct these answers, sometimes citing source pages directly — and sometimes not.
The critical distinction from traditional featured snippets is that AI Overviews can synthesise information from multiple sources into a single cohesive answer. A featured snippet highlighted one page's answer verbatim. An AI Overview might combine information from five different pages without explicitly quoting any of them — while showing a "Sources" carousel that links to some of those pages.
Not all searches trigger AI Overviews. The rollout has been selective by intent type, and Google continues to refine when and how AI Overviews appear. Understanding which queries trigger them — and which do not — is the first step in developing a sensible response strategy.
The Real Traffic Impact: What the Data Shows
Early research and industry data from SEO platforms tracking large publisher portfolios shows a consistent pattern: purely informational queries with clear, straightforward answers have seen the most significant click-through rate reduction when AI Overviews are present. Queries where the searcher wanted a quick fact, a definition, or a simple how-to answer are now often satisfied within the SERP itself.
Commercial intent queries — where someone is researching a purchase decision, comparing options, or looking for a specific service provider — have shown much less traffic disruption. Transactional queries (direct purchase or booking intent) show almost no disruption. Google's AI Overviews are not a good substitute for evaluating and purchasing products or services — they are a fact delivery mechanism, not a decision-making engine.
The practical implication: if your organic traffic was heavily weighted towards informational queries with short, factual answers, you should expect continued disruption. If your traffic is primarily commercial and transactional, AI Overviews represent a smaller direct threat — though still an important context to understand and prepare for.
Which Queries Trigger AI Overviews
Through direct testing across hundreds of query types, I have identified consistent patterns in what triggers AI Overviews. Queries most likely to trigger an AI Overview include: "how to" instructional queries, definitional questions ("what is X"), comparison questions ("X vs Y"), and multi-part questions that require synthesising information from multiple sources.
Queries least likely to trigger AI Overviews include: brand name searches, local service searches with map pack results, navigational queries (people searching for a specific website), and highly commercial or transactional queries with clear purchase intent. YMYL (Your Money, Your Life) queries in health, finance, and legal areas appear to be triggering AI Overviews with added caution — Google is particularly careful about AI generating medical or financial advice.
How to Get Cited in AI Overviews
Being cited as a source in an AI Overview is genuinely valuable — it places your brand and a link to your content prominently above the standard organic results for queries you care about. Here is what the evidence shows about how to earn those citations.
Rank well already: The clearest predictor of being cited in an AI Overview for a given query is already ranking in the top 10 organic results for that query. AI Overviews appear to draw heavily on already-authoritative pages. This means the foundational work of traditional SEO — creating authoritative, well-structured content that earns strong organic positions — remains the best path to AI citation as well.
Structure content for direct answer extraction: AI systems extract answers more reliably from content that uses clear heading structure, concise definition statements at the start of sections, numbered and bulleted lists for process-oriented content, and FAQ sections with direct question-and-answer formatting. These structural elements make it easier for AI systems to identify and extract the relevant portion of your content as an answer.
Answer the full query, not just the headline question: AI Overviews frequently synthesise answers to complex queries by pulling from pages that each answer part of the broader question. Comprehensive content that addresses the full scope of a topic — including related questions, common follow-ups, and contextual nuances — is more likely to be cited than narrowly scoped content that only answers the main keyword question.
Answer Engine Optimisation (AEO) Principles
Answer Engine Optimisation refers to the practice of structuring content specifically to be extracted and used as answers by AI systems — both Google's AI Overviews and external AI tools like ChatGPT and Perplexity that increasingly draw on web content for their responses.
Concise, direct answer first: For any question your content addresses, provide the most direct possible answer in the first 1–2 sentences following the question heading. Elaborate with supporting detail afterwards. AI systems and users alike appreciate content that respects their time by leading with the answer rather than burying it.
Schema markup as structured data signal: Implement FAQ schema on any page that addresses multiple questions. How-to schema on instructional content. Article schema with clear datePublished and dateModified. These structured data signals help AI systems understand your content's intent and recency — both factors in whether your content is selected as an answer source.
Conversational, natural language: AI language models are trained on natural human language. Content written in conversational, direct language that matches how people actually speak and ask questions tends to align better with AI training patterns than keyword-stuffed or overly formal content. Write for humans first — AI systems are increasingly good at identifying content written primarily for search bots rather than people.
Cover related questions systematically: Use tools like People Also Ask data, Answer the Public, and your own Search Console query reports to identify all the sub-questions people ask around your core topic. Structure your content to address these directly. This comprehensive question coverage is what positions your content as an authoritative resource rather than a narrow keyword-match page.
E-E-A-T Signals Google AI Trusts
Google's AI systems appear to weight E-E-A-T signals heavily when selecting which pages to cite in AI Overviews — even more heavily than traditional organic ranking might suggest. This makes intuitive sense: Google is staking its AI's credibility on the sources it cites, giving it strong incentive to preferentially cite sources it has high confidence in.
Experience signals: First-person accounts of direct experience — "I tested this," "in my work with clients," "based on 8 years of doing this daily" — are Experience signals. They distinguish content written by someone who has actually done the thing from content aggregated from other sources. These signals are increasingly difficult to fake convincingly and consequently increasingly valued.
Expertise signals: Demonstrated subject matter expertise through credentials, track record, specific technical knowledge, and domain-specific vocabulary. For professional service providers, this means making your qualifications, experience, and specific areas of expertise explicit and prominent — on your About page, in your author bio on every piece of content, and in the content itself through specific, expert-level insights rather than surface-level summaries of commonly known information.
Authoritativeness signals: Who else links to you and references you as an authority? Third-party validation — editorial backlinks from reputable publications, mentions in industry media, professional association memberships, speaking engagements — all contribute to authoritativeness signals that Google's systems can evaluate.
Trustworthiness signals: Transparent authorship (named authors with professional profiles), clear publication dates and update histories, accurate and verifiable information, a legitimate business with real contact details, and positive reviews. For YMYL topics especially, trustworthiness signals are non-negotiable for AI consideration.
Protecting Organic Traffic in an AI Search World
Beyond optimising for AI citation, the most defensible traffic strategy in 2025 involves diversifying away from purely informational content toward the query types AI Overviews do not disrupt effectively.
Commercial comparison content: In-depth comparison articles, versus pages, and alternative pages continue to drive strong organic traffic and lead conversion because AI Overviews cannot fully replicate the depth of evaluation a genuine expert comparison provides. These pages also tend to attract clicks even when AI Overviews are present because searchers making purchase decisions want to verify and explore, not just receive an AI summary.
Original research and data: Unique data that does not exist anywhere else on the web cannot be synthesised by AI from other sources. Publishing original surveys, case study data, and proprietary research creates content that can only be cited, not replicated, by AI systems — making it inherently defensible.
Experience-led content: Detailed, first-person accounts of doing complex things — case studies, project recaps, personal testing and reviews — are both E-E-A-T goldmines and AI-resilient content formats. They rely on experience that AI systems cannot credibly replicate or substitute for.
Local and brand content: Queries with local intent and brand-specific queries are not being disrupted by AI Overviews in any meaningful way. For local businesses especially, the Local Pack and local SEO signals remain the critical battlefield — and they are as human-reviewed and competitive as ever.
The fundamentals of SEO — authoritative content, clean technical infrastructure, strong E-E-A-T signals, and relevant backlinks — are what earns citations in AI Overviews and what maintains rankings in traditional organic results. The strategy has not changed; the emphasis on certain elements has intensified. Lean into that.
If you want help adapting your specific content strategy for the AI search era, my AI SEO service is designed exactly for this challenge. Or book a free consultation to discuss your situation directly.
Mani Pathak specialises in AI SEO strategy and has been tracking the impact of Google's AI Overviews since the initial SGE rollout. He advises clients across industries on adapting to AI-driven search changes.