Loading…
Loading…
We analysed 1,200 keywords across 14 UK SME websites. The real impact of AI Overviews is more nuanced than the headlines — and there are clear moves you can make right now.
Every SEO conference in 2026 has the same keynote: AI Overviews are destroying organic traffic. And they're not wrong — for certain query types, click-through rates have collapsed. But after managing SEO for 14 UK SME websites over the past 12 months, we've found the story is significantly more nuanced than 'Google is stealing your traffic'.
Between October 2025 and March 2026, we tracked 1,200 keywords across these 14 sites — spanning B2B services, eCommerce, SaaS, and professional services. We measured which queries triggered AI Overviews, what happened to CTR when they appeared, and critically, which content types and strategies held their ground or even grew.
The headline finding: AI Overviews appeared on 43% of the keywords we tracked. But the traffic impact ranged from catastrophic (–74% CTR) to literally zero — depending entirely on query type, content format, and how the page was structured. The difference between winning and losing in this new reality isn't luck. It's strategy.
Let's start with the raw data. Of 1,200 tracked keywords across our client portfolio, 516 (43%) triggered an AI Overview in UK search results as of March 2026. That's up from roughly 28% when we started tracking in October 2025 — Google is expanding AIO coverage aggressively, adding approximately 3 percentage points per month.
For the 516 keywords with AI Overviews, average CTR dropped from 3.1% to 0.8% — a 74% decline. That's the scary number everyone quotes, and it's accurate. But it's an average, and averages hide everything interesting.
When we segmented by query intent, the picture sharpened dramatically. Informational queries ('what is X', 'how to X', 'X explained') lost 81% of their CTR. These are the queries where AI Overviews provide a complete answer — users get what they need without clicking. For SME blogs built on 'ultimate guide' and 'explainer' content, this is an extinction event.
Commercial investigation queries ('best X for Y', 'X vs Y', 'X pricing UK') lost only 23% of CTR. AI Overviews appear on these queries, but users still click through because the Overview can't replicate subjective evaluation, price comparison, or real user experience. Transactional queries ('buy X', 'X near me', 'hire X') were barely affected — 8% CTR decline, mostly from the Overview pushing organic results slightly lower on mobile.
The most surprising finding: navigational and branded queries actually gained 4% CTR when an AI Overview was present, because the Overview often reinforced the brand and provided a direct link. If people are searching for your brand name, AI Overviews help, not hurt.
The businesses suffering most from AI Overviews share a specific content profile. They built their organic strategy around informational blog content — 'What is digital marketing', 'How to improve your website SEO', 'Guide to email marketing for small businesses'. Two years ago, this was solid SEO advice. Today, it's building on quicksand.
Across our 14 client sites, we identified 187 pages that lost more than 60% of their organic traffic between October 2025 and March 2026. The pattern was consistent: 89% of these pages targeted informational queries, 78% were longer than 2,000 words (the 'comprehensive guide' format), and 91% had no original data, proprietary research, or interactive elements.
The content graveyard isn't random. Google's AI can synthesise a satisfactory answer for any question that has a factual, consensus-based response. If your content is essentially restating publicly available information in a well-structured format, you've been replaced by a machine that does it faster and presents it directly in the SERP.
But here's the flip side — 134 pages across the same 14 sites actually grew traffic during the same period. These pages shared three characteristics: they contained original data or proprietary insights (62%), they targeted comparison or evaluation queries (48%), or they included interactive elements like calculators or configurators (31%). Some pages had two or all three of these characteristics, which is why the percentages exceed 100%.
After six months of data, we've identified five query categories where organic CTR remains strong despite AI Overview presence. These aren't theoretical — they're drawn directly from our client data.
First: comparison queries. 'HubSpot vs Pipedrive for small teams', 'Shopify vs WooCommerce UK costs', 'SEMrush vs Ahrefs 2026'. AI Overviews attempt these but produce flat, uncommitted comparisons that satisfy nobody. Users want an opinion. They want someone who's actually used both tools to tell them which one is better for their situation. Our comparison pages retained 82% of their pre-AIO click-through rate.
Second: queries with a local or regional dimension. 'Best accountant for contractors in London', 'SEO agency Manchester reviews', 'commercial electrician rates Bristol'. AI Overviews struggle with hyperlocal knowledge, and users inherently trust local recommendations less when they come from a generalist AI. CTR retention: 79%.
Third: experience-based queries. 'What it's actually like migrating from WordPress to Shopify', 'Honest review of Google Performance Max after £20K spend', 'Lessons from scaling a B2B SaaS from 10 to 100 customers'. These require lived experience that an AI can't fabricate convincingly. The E-E-A-T signal here is genuine experience, and users can smell the difference. CTR retention: 76%.
Fourth: current-data queries. 'Average Google Ads CPC UK 2026', 'Instagram engagement rate benchmarks Q1 2026', 'UK eCommerce conversion rate by industry'. AI Overviews draw from training data that's months or years old. If your content has data from this quarter, you have something the AI doesn't. CTR retention: 71%.
Fifth: tool and calculator queries. 'ROAS calculator', 'marketing budget planner', 'SEO audit tool'. These require user input and deliver personalised output — something an AI Overview structurally cannot do. CTR retention: 94%. This is the single most resilient content category we've measured.
The strategic implication is clear, but uncomfortable for many SMEs: the era of 'answer the question' content is over. Google's AI does it better, faster, and without requiring a click. The new content game is about providing something the AI can't — judgement, experience, proprietary data, and interactive utility.
For our clients, we've implemented what we call the 'Judgement Layer' content framework. Every piece of content must pass a simple test before we create it: does this require human judgement, original data, or user interaction that an AI Overview cannot replicate? If the answer is no, we don't write it. Period.
Practically, this means shifting from 'What is conversion rate optimisation' to 'We A/B tested 23 landing pages for UK B2B sites — here's what actually moved conversion rates'. From 'How to choose a CRM' to 'The CRM decision matrix: 6 questions that predict which platform you'll still be using in 3 years'. From 'SEO best practices' to an interactive SEO health checker that analyses the user's own URL.
The content volume has dropped — we're publishing roughly 40% fewer posts per month than a year ago. But the traffic per post is 3.2x higher, the conversion rate is 2.7x higher, and the content has a longer shelf life because it's not competing with an AI that can generate generic advice infinitely.
One critical mindset shift: stop thinking about 'ranking for keywords' and start thinking about 'being the source the AI cites'. For queries where AI Overviews are inevitable, our goal shifted to appearing in the AIO citation links. Pages cited in AI Overviews still get clicks — typically 15–22% of the pre-AIO CTR, which is better than being on page 2. This requires structured data, clear entity markup, and content that the AI recognises as authoritative on the specific topic.
One of the clearest patterns in our data: pages with comprehensive structured data markup were 2.4x more likely to be cited in AI Overviews than equivalent pages without it. And cited pages retained significantly more traffic than non-cited pages that lost their traditional organic positions.
We're talking about more than basic Schema.org Article markup. The pages that performed best had specific, granular structured data: FAQPage schema with genuine questions from their customer base, HowTo schema with clear step breakdowns, Product schema with real pricing and availability, and LocalBusiness schema with service area definitions.
For our B2B clients, we added Organisation schema with detailed department and service descriptions, ProfessionalService schema with area served and price range properties, and Review schema aggregating genuine client testimonials. Google's AI uses this structured data to understand what a page actually contains, and pages with rich markup are more likely to be selected as citation sources.
The ROI of structured data implementation has shifted from 'nice for rich snippets' to 'essential for AI Overview survival'. Across our 14 sites, implementing comprehensive structured data on existing pages recovered an average of 18% of the traffic lost to AI Overviews — not by bypassing the Overview, but by getting cited within it.
Based on six months of data, here are the five highest-impact actions for UK SMEs facing AI Overview traffic loss.
Move 1: Audit your content against AI Overviews. Search your top 100 keywords in incognito. Categorise each page as safe (no AIO), cited (AIO present, you're cited), or displaced (AIO present, you're not cited, traffic falling). This gives you the battle map. Budget one full day for this — it's the most important day of SEO work you'll do this year.
Move 2: Kill or consolidate your informational dead weight. If a page targets a purely informational query, gets an AIO, and you're not cited, that page is consuming crawl budget for zero return. Redirect it to your most relevant commercial page. We consolidated 187 dying pages into 43 stronger, commercially-focused pages across our client sites — and total organic value increased despite losing total page count.
Move 3: Add original data to everything. Survey your customers. Benchmark your industry. Analyse your own campaign data. Anonymise and publish it. 'The Average UK SME Marketing Budget in 2026: Data from 200 Businesses' is something AI Overviews can't generate from training data. These pages become citation magnets and attract backlinks from industry publications. Three of our data-led posts were cited by Marketing Week within a month of publication.
Move 4: Build one interactive tool. A budget calculator, a readiness assessment quiz, a cost comparison tool — anything that takes user input and delivers personalised output. Our clients' calculator pages have an average conversion rate of 8.3% versus 2.1% for blog posts. They're AI-proof by design, and they generate leads directly. Even a simple spreadsheet-style calculator embedded on a page outperforms any written guide.
Move 5: Implement comprehensive structured data now. Not just Article schema — FAQPage, HowTo, LocalBusiness, ProfessionalService, Product, and Review schema where applicable. Pages with full structured data are 2.4x more likely to be cited in AI Overviews. This is the fastest, most cost-effective action on this list — it takes a developer a day or two and the impact is measurable within weeks.
Google has been transparent about expanding AI Overviews, and the trajectory is clear. Based on our tracking, AIO coverage is growing at approximately 3 percentage points per month in the UK. At this rate, more than 60% of non-brand, non-navigational queries will have an AI Overview by October 2026.
We're also seeing AI Overviews becoming more sophisticated. Early Overviews were text-heavy summaries that occasionally cited sources. Recent Overviews include structured tables, comparison matrices, step-by-step visuals, and embedded maps. Google is investing heavily in making the Overview the complete answer, reducing the need for any click at all.
The emerging counter-strategy is what we're calling 'post-click value' — making the page experience so valuable that even when a user gets an AI Overview summary, they click through for the deeper experience. This means interactive elements, downloadable tools, video walkthroughs, and personalised content that requires the user to be on your site. The blog post as a static wall of text is dying. The blog post as an interactive experience is just beginning.
For UK SMEs specifically, the local angle remains a significant advantage. Google's AI is weakest on hyperlocal content, recent data, and subjective experience — exactly the areas where small businesses have the most authentic expertise. If you're a Manchester-based marketing agency, your lived experience working with Manchester businesses is something no AI can replicate. Lean into that specificity rather than competing on generic advice.
We'll audit your top keywords against AI Overviews, identify which pages are at risk, and build a content strategy around the query types that still drive clicks — original data, comparison content, and interactive tools.