Search has changed shape. The familiar list of blue links now shares the stage with AI Overviews and a conversational AI Mode that answers complex questions in one flow. The effect is structural. Marketers face fast technology shifts, new user habits, and stricter oversight in the UK and EU. Treat this as a system problem. Understand the mechanics, the market incentives, and the legal context. Then decide where to win attention, where to win clicks, and where to win trust.
The practical stakes are clear. AI Overviews compress information at the top of the results, often satisfying need without a visit. AI Mode encourages longer, multi-step sessions that produce more queries per user. In both cases, visibility becomes multidimensional. You optimise not only to rank but to be cited, to be chosen, and to be remembered. The task is to rewire SEO for answer engines while protecting brand equity and revenue.
How Google splits intent with AI Overviews and AI Mode
Google now uses two surfaces for different parts of the journey. AI Overviews summarise answers for broad informational intent. They sit at the top of the page and cannot be disabled by users. AI Mode runs on Gemini 2.5 and turns a search into a conversation. It handles multi part prompts, breaks them into sub searches, and synthesises output with links. Early usage showed longer prompts and more follow-ups per session. For marketers, the split matters. AI Overviews serve speed. AI Mode serves depth. Strategy must account for both.
Search is bigger than Google in the answer engine era
The definition of search now includes chat interfaces where discovery starts upstream of Google.com. ChatGPT and other assistants capture a large share of “first questions”, with Google often receiving the later clicks. Meanwhile, UK market share for Google remains dominant, which makes its AI choices decisive. This dual reality forces a two-front plan. Compete for presence inside AI answers across platforms. Also strengthen the parts of your funnel that you fully control, including email, community, and direct navigation.
Adoption trends support the shift. Reports in 2024 and 2025 showed rapid enterprise uptake of generative tools and record private investment. As AI becomes routine in work and daily life, users expect direct, synthesised answers. Brands that adapt content to this expectation gain attention even when clicks are scarce.
Regulation and publisher pushback reshape incentives
AI Overviews triggered formal complaints in the UK and EU from publisher coalitions and advocacy groups. Core claims include the absence of a meaningful opt-out and the risk of traffic loss caused by summaries that satisfy intent on the page. The UK CMA has moved to apply Strategic Market Status to Google Search and search ads, enabling conduct requirements. Publisher bodies have asked for separate crawl reporting for indexing versus model training, AI performance data in Google Search Console, and clearer attribution inside AI answers. In parallel, the EU AI Act sets obligations for general-purpose models. Outcome is uncertain. Strategy should assume continued change in attribution, reporting, and opt outs.
The great traffic decoupling and why CTR is falling
The link between rank and traffic has weakened. Evidence across sectors shows stable or rising visibility with falling click through rates where AI Overviews appear. Zero-click outcomes grow because summaries meet needs on the page. Case data from publishers indicate CTR drops even in position 1. In some tests, desktop and mobile CTR fell sharply when an AI summary was present. Analytics firms have modelled large traffic losses when a top organic result moves below an AI Overview.
This is not uniform. It varies by query type and industry. What holds across categories is the new gap between “seen” and “visited”. You now optimise for impressions within the AI element, for brand recall among cited sources, and for the subset of users who still click for depth.
Where AI Overviews appear by intent
Map AI intervention to intent types.
Informational intent. High trigger rate. “What”, “when”, “where”, and “how” queries often generate multi-paragraph summaries with bullets and multiple citations. These are the main driver of zero click behaviour.
Commercial investigation. Medium to high trigger rate. Comparison queries may show tables, feature digests, and quote snippets. Users can shortlist options without leaving the page.
Transactional intent. Low trigger rate. Shopping units, ads, and direct links dominate. Summaries would slow conversion, so Google largely suppresses them.
Navigational intent. Very low trigger rate. Users seeking a specific brand or login get a direct link. AI Overviews add little value and rarely appear.
Inside the answer engine synthesis process
For complex prompts, Google fans out the query into related sub searches, collects results across subtopics, and then synthesises a single response with citations. This favours clear information architecture and consistent signals. Well structured pages with strong schema markup are cheaper to parse and easier to cite. Analyses of long-form answers show broad source diversity for complex topics. Featured snippets and AI Overviews often coexist because both target the same informational need.
Fun fact: When you structure a page with crisp headings, short answer boxes, and validated schema, you reduce the model effort needed to extract facts, which raises your odds of citation even if your domain is smaller than incumbents.
Strategic trigger matrix for planning
Use a simple matrix to decide aims and tactics.
Informational. Aim to earn citations and brand impressions. Tactics: short direct answers near the top, clear lists, FAQs, and visible authorship. Strengthen E E A T signals to increase trust.
Commercial investigation. Aim to win the click and the citation. Tactics: build “unsummarised value” through proprietary data, interactive tools, and sharp comparisons that reward a visit.
Transactional. Aim to secure conversion. Tactics: fast product pages, strong CRO, and pay close attention to shopping units and local packs.
Navigational. Aim to be found fast. Tactics: Reinforce brand recall through PR and community, so more users search for you by name, where AI rarely interrupts.
Pillar 1 Foundational authority with E E A T
Experience. Show real use. Include original photos, videos, logs, and process details tied to people. Add outcome data and constraints faced in practice.
Expertise. Demonstrate depth and correctness. Use precise terms, cite reputable sources, and explain trade-offs. Tie content to a named author with credentials and a verified profile.
Authoritativeness. Earn recognition from peers. Seek high-quality links and mentions from respected publications and institutions in your niche.
Trustworthiness. Make the site safe and transparent. Use HTTPS, clear contact paths, privacy and editorial policies, and visible corrections. Keep facts consistent across brand profiles.
These signals matter to users and machines. They help the answer engine determine whether your content is safe to quote and recommend.
Pillar 2 Technical SEO as the language of machines
Technical clarity removes ambiguity. JSON-LD schema tells systems what an entity is and how fields relate. Prioritise:
Article for authorship and dates.
Product for model, price, stock, and ratings.
FAQPage for Q and A blocks.
Review for structured feedback.
Organisation and Person for entity identity.
VideoObject and ImageObject for media discovery.
Validate markup with official testing tools and keep on-page content in lockstep with structured data. Then fix crawl and index basics. Use clean URLs, consistent headings, tight internal linking, and canonical tags. Keep pages fast and stable. AI systems prefer content that is easy to parse, cache, and reuse.
Pillar 3 A bifurcated content strategy
Decide the goal before you write. Either be easy to cite or be hard to compress.
Citable content. Target broad informational queries where clicks are sparse but impressions are large. Open with a 40 to 60 word direct answer to the core question. Use scannable sections that mirror common queries. Add short checklists and FAQs. The brand win is association. Measure by impressions within AI elements and growth in branded search.
Unsummarised content. Target high-value commercial investigation and transactional queries. Offer value that a summary cannot replicate. Use proprietary datasets, calculators, configurators, diagnostic quizzes, live comparison tables, and case studies with non obvious findings. Personalise where possible so the page requires user input. Measure by leads, assisted conversions, and lifetime value.
Pillar 4 Building an unassailable brand
When an AI Overview lists several sources, users choose the one they recognise. Brand trust becomes the deciding factor. Treat PR, community, events, and thought leadership as core SEO inputs. Grow navigational demand so more sessions start with your name. That traffic is resilient because AI rarely intervenes on branded queries. Align design, tone, and claims across all surfaces to avoid trust gaps.
Pillar 5 Risk management and content integrity
Generative systems can hallucinate. Protect the brand.
Human in the loop. Edit and fact check everything. Use subject experts for sensitive topics. Publish corrections promptly.
Proactive monitoring. Track brand mentions on the open web and in assistant outputs where possible. Document issues and request fixes through formal channels.
Technical controls. Where a page is frequently misquoted, limit previews using snippet directives such as nosnippet, data nosnippet, and max snippet. Balance the loss of preview reach against the gain in accuracy.
Measuring what matters in AI features
Abandon rank as the sole metric. Build a model that measures exposure, selection, and business impact across AI surfaces.
Exposure. Impressions within AI Overviews and related elements.
Selection. CTR when cited, share of citations on a topic, and scroll depth after click.
Impact. Branded search lift, assisted conversions, and downstream engagement.
Attribute by topic, not only by keyword. Users now move through conversational chains. A single article can appear multiple times as a session unfolds.
New GSC workflow for AI Overview analysis
Use Google Search Console to isolate AI appearance.
Filter by Search appearance set to AI Overview.
Export queries and pages to identify the topics where you are cited.
Compare CTR for cited queries against site averages to quantify the impact of summaries on your content.
Group queries into topic clusters. Track impressions, clicks, and CTR for each cluster to understand where you win citations and where you win visits.
Now build actions. For clusters with high impressions and low CTR, improve title clarity, tweak meta snippets, and raise brand familiarity around those topics through PR and partnerships. For clusters with strong CTR, expand related articles to capture more of the conversational chain.
Beyond the click measuring brand lift and citation value
Clicks do not capture the full value of a citation. Treat AI Overview exposure like an upper funnel impression and measure brand response.
Search lift. Compare branded query volume between exposed and control cohorts to estimate the increment from citations.
Brand lift. Use surveys to test whether exposure within an AI element increases recognition, consideration, and intent.
Memory hooks. Add distinctive elements to cited pages, such as concise frameworks, named methods, or original visuals, to raise recall.
Use these inputs to decide whether to invest in citable content for a topic or to switch to unsummarised assets aimed at conversion.
Solving attribution in AI Mode
AI Mode treats each follow-up as a fresh query. That multiplies opportunities for a single strong page to appear several times in one session. Plan content for the whole dialogue, not a single question. Build deep topic clusters with internal links that answer likely follow-ups in sequence. Include short answer boxes within sections to make extraction easy. When performance is strong for one query in a chain, add pages that answer the following two links in the chain and cross link them.
Content clustering for conversational journeys
Design clusters around user jobs, not just keywords. Interview customers. Read support tickets, sales notes, and community posts. Capture the language they use and the order in which they ask questions. Turn that into a path of articles, tools, and case studies. For each node in the path, determine whether it is citable or unsummarized. Keep navigation clean so users can jump forward or deeper. Use interlinking that reflects the path users take in conversations, not only the site tree.
Action checklist for GEO in 2025
Define primary topics and choose citable versus unsummarised aims per topic.
Upgrade authorship with full bios, credentials, and consistent entity markup.
Add JSON LD for Article, Product, FAQPage, Review, Organization, Person, VideoObject, and ImageObject where relevant.
Refactor pages to open with a 40 to 60 word direct answer where intent is informational.
Ship at least 3 interactive or data driven assets for your highest value commercial topics.
Harden trust signals with clear policies, visible contacts, and a published corrections process.
Set up GSC AI Overview filters and build topic level dashboards.
Measure branded search lift after major citation wins.
Expand topic clusters to answer 3 to 5 linked follow-ups per core query.
Test snippet controls on pages that are repeatedly misrepresented.
Conclusion thriving in the answer engine era
Search is now an answer engine. AI Overviews compress information at the top. AI Mode turns sessions into conversations. Ranking alone does not guarantee traffic. The path to growth involves winning trust, earning citations, and creating assets that compel a visit when it matters. The five pillars of Generative Engine Optimisation provide the frame. Prove human experience and expertise. Speak clearly to machines. Split content aims by economics. Build a brand users choose when several sources appear. Manage risk with process and controls. Then measure across exposure, selection, and impact at the topic level.
Users will reward clear thinking and practical value. Give them both. Treat each query as the start of a dialogue that you can shape. When your content answers the next question before it is asked, you become the default choice in a crowded page. That is how you win traffic and trust in 2025. In simple terms, consistent work builds predictable returns, and as the proverb says, “Slow and steady wins the race.”
