Long gone are the days when most Google searches meant a click through to a website. The search landscape is undergoing a seismic shift thanks to AI-driven results and large language models (LLMs).
If you’re an eCommerce or marketing manager watching your Search Console graph turn into an “open-mouthed crocodile” – impressions rising sharply while clicks decline – you’re not alone. More and more search queries in the US and UK now trigger an AI-generated overview at the top of the results, siphoning attention away from the traditional links and decreasing click-through rates. This is likely to spread to more countries and become more prevalent in the ones it’s already seen in.
This post breaks down what’s changing, what it means for your business, why SEO needs a technical refresh, and what you should do now to adapt.
AI Answers Are Everywhere: Google and Bing are increasingly blending AI-generated answers into search results. Google’s new Search Generative Experience (SGE) can summarise answers at the top of the page, often with cited sources. These AI Overviews started with niche questions and long-tails but are now appearing for a wide range of queries, even general product searches.
One recent analysis found Google’s AI Overview feature on about 18.9% of all US search keywords right after a major rollout in March 2025. That figure has only grown since (more than doubling after the March core update), meaning nearly one in three searches now shows an AI summary before the organic results.
Our research for a client that sells technical equipment to manufacturing and research facilities has found that there is the presence of an AI Overview on around 40% of the page 1 results in the UK and US for their keywords. We found that a lot of these results also had a featured snippet, which pushed the organic results even further down the page, particularly on mobile.
In parallel, Bing’s Chat mode (powered by GPT-4) is answering questions directly in its results. And outside of the traditional engines, standalone LLM chatbots like ChatGPT have amassed hundreds of millions of users (ChatGPT’s weekly active user count grew 8× to over 800 million by early 2025).
The upshot: more people are getting their answers within the search interface, without clicking through.
Zero-Click Searches Are Soaring: Even before generative AI, the trend was toward zero-click searches, where the answer is given on the results page. Google’s featured snippets, knowledge panels, maps, and other rich results were already providing immediate answers. In 2024, an estimated 60% of all US Google searches ended without a click. Now, AI overlays are supercharging this trend.
Google’s AI summaries often provide a comprehensive answer drawn from multiple websites, sometimes listing sources only as small icons. Users can get what they need at a glance, so why click?
In one analysis across 300,000 searches, top organic results lost 34.5% of their clicks when an AI answer was present.
In other words, for any given query, if Google’s AI can satisfy the intent, a large chunk of would-be visitors never leave the SERP.
Double Impressions, Half the Clicks: Paradoxically, your pages might be getting seen more than ever, yet clicked less. How? With AI answers, content that ranks well can appear twice on page one – once in the traditional results and again as a cited source in the AI summary. That means two impressions for one query.
Many AI Overview citations are pulled from pages already ranking in the top results (one study found 75% of cited links were from the top 1–2 page positions). So if you’re ranking for a query, you also have a good chance of being featured in the AI blurb, effectively doubling your impressions.
At the same time, because the user might get their answer from the overview, they’re less likely to click either result. This means clicks and impressions have become decoupled from one another; one can grow while the other declines. This is a major change to a long-standing and fundamental rule of SEO. It’s why your Search Console graph might show record-high impression counts but a flat or falling click line. The rules of engagement in SEO are being rewritten in real time – visibility no longer guarantees traffic.
AI as the Default Experience: Both Google and Microsoft are betting on AI as the future of search. Google’s AI Overviews (or the recently launched in the US, “AI mode”) are in line to become the default search interface sooner rather than later. Instead of the ten blue links, with a scattering of SERP features, users would see an interactive Q&A-style result by default, with traditional links playing a secondary role.
This means the global organic traffic slump could persist or deepen as these AI answers become the norm. In short, search is turning into a conversation, and websites have to find new ways to be heard in that conversation.
Impressions Up, Clicks Down (Yes, It’s Real): If your organic traffic report looks puzzling – rising impressions but fewer clicks – you’re witnessing this AI effect firsthand. There have been observed cases of impressions surging over 50% while clicks dropped by double digits in the same period. Your content is still showing up on Google, sometimes even more than before, but users aren’t clicking through as often. The metrics below illustrate this pattern for one of our clients, often called the “crocodile trend,” where the gap between impressions and clicks widens over time:
For eCommerce brands, this can be alarming. Fewer clicks mean fewer opportunities to bring customers to your product pages or content. However, those impressions are not worthless – they indicate your brand or content is being seen as part of AI-generated answers.
In practical terms, a shopper might ask an AI, “What’s the best running shoe for trail marathons?” and see an AI summary that mentions your brand’s trail shoe in the answer. They might not click immediately (no need if the AI already pulled your product info), but your brand has been introduced into their consideration set. We’re effectively moving into an era of “brand impressions without clicks.”
The New Customer Journey: AI-assisted search is compressing the classic marketing funnel. Users get their research done directly on the SERP. By the time they do visit a site, they’re often further along in the decision process than in the traditional funnel. For eCommerce, this means the traffic you do get from organic search may be lower in volume but higher in quality. Industry data suggests that when users arrive via an AI-powered search (like clicking a link in an AI overview or coming from Bing’s chat results), they convert significantly better. One study found the average LLM-driven visitor is about 4.4× more likely to convert than a traditional search visitor.
So by the time someone clicks through an AI summary, they’ve essentially been pre-qualified by all the upfront information.
Bottom line: You may see fewer bites, but the ones you do get are hungrier.
Brand Visibility Without Traffic: In the AI era, brand awareness can grow even as click metrics shrink. You might notice an uptick in direct traffic or searches for your brand name that doesn’t line up neatly with your Google click count. This is a clue that people are discovering you via AI recommendations and then visiting later.
For eCommerce businesses, especially those selling niche or considered products, this kind of exposure is gold. It’s like getting mentioned in a personal recommendation engine.
However, it’s also harder to track and attribute – your analytics might show it as a direct visit or “organic” with the query (not provided). The traditional funnel attribution will be muddied until reporting catches up with search technology and trends.
Competitive Landscape Shake-Up: AI-driven search can level the playing field in unexpected ways. The sources that an AI chooses for answers aren’t always the ones with the highest traditional SEO rankings. They might favour a very specific paragraph from a smaller site because it answers the question precisely.
Competitors who have weaker classic SEO might still snag a spot in AI results due to offering a uniquely relevant tidbit of information. If your brand has relied on dominating the SERPs, be aware that LLMs don’t read the name on the building – they read the content. For eCommerce, this means a smaller competitor with an excellent, informative guide could get picked as an authority in an AI summary over a big-brand site. Conversely, if you produce the most useful content on a topic, you might punch above your weight in AI visibility. It’s a chance to win ranking share without necessarily winning the click or the #1 ranking.
Pressure on Content and Conversion: With fewer but more qualified clicks, every visit counts more than ever. If AI search is skimming off casual informational visitors, the ones who do land on your site are likely highly interested. You’ll want to make sure your site is ready to convert those high-intent visitors efficiently. That means optimising landing pages, tightening up your checkout process, and ensuring product pages answer any remaining questions (availability, shipping, reviews, etc.).
Get in touch with us if you’d like to discuss improving the conversion rate of your eCommerce website.
Also, consider that if users saw some of your content in an AI snippet, they arrive with context. The landing page should match that context so they feel continuity. For example, if the AI overview mentioned “free returns at [YourStore] for trail shoes,” the page they clicked had better highlight that policy clearly.
Organic SEO is shifting from a game of volume to a game of impact. Fewer visits, but each visit carries more weight for your revenue, which raises the stakes on providing a seamless on-site experience.
In this new landscape, sticking to “old school” SEO tactics isn’t just ineffective – it’s risky. Now is the time to revisit your site’s technical foundations and adaptability. Here’s why a technical refresh is crucial:
LLMs Lag Behind Live Web: Traditional search engines like Google pride themselves on fresh results – they crawl and index pages constantly, so when you update your site, Google Search usually reflects changes quickly. LLMs, on the other hand, operate on delayed knowledge. A chatbot like ChatGPT might be trained on data that’s months or even years old. Even Google’s own AI summaries, while based on the live index, may occasionally surface content that’s outdated if your site’s content has changed recently.
The consequence? A higher chance that an AI is citing or sending users to out-of-date pages or broken links. If you’ve removed or moved a page that an AI system “remembers,” users could get a 404 error when they try to click through. This is a bad look for your brand and a lost customer opportunity. Technical SEO best practices – like setting up proper redirects and maintaining stable URL structures – are more critical than ever. If an old URL of yours is floating around in AI training data, make sure it at least points to a relevant new page instead of a dead end.
Site Health = AI Health: Along with fixing broken links, you need to keep your site crawlable, fast, and structured. Remember, AI systems gather their info by crawling the web (Google’s SGE uses Google’s index; other LLMs use data from Common Crawl and similar). If your site has poor crawlability – say broken sitemaps, slow response times, or restrictive robots.txt – you might be invisible to these AI content harvesters. A technical audit should check that all your important content is easily accessible to crawlers (both search engine bots and the generic ones like CommonCrawl that feed AI models). Page speed and performance matter too. While an AI summary might not care if your site is slow (since it’s not loading your live page when generating an answer), user experience does. If a user clicks from an AI result to your site, a sluggish page will prompt them to bounce right back to the answer that got them there. Moreover, Google’s algorithms – including whatever powers AI overviews – still use core web vitals and other quality signals to judge if a page is worth showcasing. In short, double down on the fundamentals: fast load times, mobile-friendly design, clean code, and no glaring crawl errors.
Fresh Content, Consistent Data: It’s not just your site’s code that needs a refresh – your content strategy does too, in a very technical sense. AI answers prioritise accuracy and relevance. Google’s representatives have touted their fresh index as an advantage over systems like ChatGPT, which might hallucinate outdated information. To leverage that, you must ensure your content is up-to-date and factually correct.
This might mean more frequent content updates and pruning of stale pages. For example, if you have old product pages lingering (out-of-stock or outdated models), update them with notices or redirects. An AI overview could easily surface an old price or spec from an outdated page if it’s still live and not marked. Don’t feed the AI incorrect data. Structured data (schema) also plays a role here: by marking up product information, FAQs, reviews, etc., you make it easier for search engines (and indirectly LLMs) to extract the correct details from your site. Google’s AI overviews often draw on schema-backed info for things like product specs or business information. Make sure yours is correct and consistently formatted across the site.
Semantic Search and Multi-Vector Matching: Behind the scenes, search engines are getting smarter about how they retrieve and rank information. Google’s recent MUVERA algorithm (a multi-vector retrieval method) is a great example. Rather than relying solely on keyword matches, modern algorithms use vector embeddings to understand context and similarity between queries and content. In practice, this means Google is better at matching the intent of a query with pages that truly answer it, even if the wording is different.
For SEO, this is a clear signal: stop obsessing over exact-match keywords. Ensure your content comprehensively covers the topic and related subtopics. If someone searches “corduroy jackets men’s medium,” a context-aware system like MUVERA will surface the page that sells men’s medium corduroy jackets, not just a blog that happens to mention those words. Your SEO strategy should adapt by focusing on semantic relevance and completeness. It’s a technical shift in how we optimise content – call it the move from keyword SEO to concept SEO.
Audit your content: Are you providing depth on topics, answering follow-up questions, and using natural language that matches what users might ask? This matters not only for ranking in normal search but also for getting picked up in AI answers (which look for the piece of text that best addresses the question).
Measuring the Invisible: Another aspect of the technical refresh is how we measure SEO success. The old metrics (rankings, clicks, even page views) don’t tell the full story anymore. Google has just started to include AI-driven impressions and clicks in Search Console data, but they lump them in with regular search stats, so they can’t be isolated from GSC metrics. There’s currently no easy way to separate how many times your site was shown in an AI snippet versus a normal result. And for purely conversational platforms like ChatGPT, there’s no native analytics.
This means SEO teams need to get creative with measurement. It might involve using third-party tools or scripts to detect traffic from AI sources, checking your server logs for unusual referrers (e.g., chat.openai.com), or simply monitoring brand mention volume online. It’s a more technical challenge than before to prove the value of SEO. But it’s one we have to tackle: you may need to report new KPIs like “AI impressions” or track correlation between higher impressions and direct traffic. In essence, we must adapt our analytics stack to capture the full impact of organic visibility in the AI age.
If you haven’t yet, familiarise yourself with the new Search Console filters and Bing Webmaster Tools data – ensure you’re looking at impressions (not just clicks) to gauge your real reach. And accept that some of that reach will be “dark” – i.e. not directly trackable – requiring you to connect the dots with indirect evidence.
It’s not business as usual, but the good news is there are concrete steps you can take to thrive in this new environment. Here are specific recommendations to adjust your SEO strategy starting today:
Embrace the New Metrics: Redefine how you evaluate SEO performance. Don’t just report on clicks, and it’s probably time to sunset keyword rankings as a KPI. Start tracking impressions, especially those from AI-infused results. In Google Search Console, keep an eye on queries where your content got impressions but few clicks – those might be AI Overview cases. Similarly, use Bing Webmaster Tools to monitor your Bing visibility. Understand that a rise in impressions with flat clicks is the new normal, and educate your stakeholders that this still signals success (brand exposure!). If you see direct traffic or brand-search volume increasing, highlight that as an outcome of unseen AI referrals. In short, broaden your definition of “SEO ROI” to include visibility and assistive value, not just last-click traffic.
Audit Your AI Visibility: Just as you’d audit your search rankings, audit where and how your brand appears in AI answers. This is a new kind of audit. Try queries in Google’s SGE (for those enrolled in it) related to your products or content – are you cited? Do the same on Bing’s chat and even on ChatGPT (with plugins or browsing enabled). Tools are emerging for this (Semrush and Ahrefs have beta features that list AI citations), but you can start with manual checks on key questions in your niche. The goal is to get a handle on LLM visibility: which topics or questions does the AI associate with your brand? Which ones are dominated by competitors? This will reveal content opportunities where you can create a better answer and potentially get featured. It’s akin to rank tracking, but for answer boxes and AI snippets.
Double Down on Technical SEO: Conduct a thorough technical SEO audit with an eye toward resilience in an AI-driven world.
Think of this as future-proofing your site: the cleaner and more robust your infrastructure, the more confidently you can be integrated into evolving search features.
Make Content AI-Friendly (Without Losing Quality): To get picked up by AI summaries, your content needs to be easily digested by algorithms. This doesn’t mean dumbing it down; it means structuring it for clarity. Use clear headings and subheadings that might match question phrasing (e.g., “How to Choose a Kayak Size?” as an <h3> if you want to answer that). Bullet points and concise paragraphs help, as they are easily quotable. Directly answer common questions in your copy – for example, include a brief Q&A in your product pages or an FAQ section with obvious questions. This “answer snippet optimisation” increases the chance that an AI will pull your text as the answer. Also, incorporate schema markup for FAQs, How-To steps, product details, etc., which can feed not just search features but also train AI models with correct info. The key is to serve up facts and answers in bite-sized chunks. Think of your content as a data source for these AI systems. The easier you make it for them to extract a correct, stand-alone statement, the more likely you’ll be featured (and the less likely they’ll misconstrue something). Just be sure to maintain your brand voice and depth – we still want real humans to find value when they click through.
Keep Content Fresh and Relevant: As noted, LLMs can suffer from stale data. Meanwhile, Google’s AI is leveraging very current index information. Leverage that to your advantage: refresh your content regularly. Update old blog posts with new insights or 2025 statistics, so that if an AI pulls info from it, it’s up-to-date. Remove or mark outdated content – e.g., add “(2024)” in titles of older annual reports, or archive pages that are no longer applicable. An AI overview might not always discern the timeliness of content unless you make it obvious.
For eCommerce, ensure product availability and pricing info is current on your pages (sounds basic, but if an AI cites your price or “in stock” status from last crawl and it’s wrong, customers lose trust). Monitor trending questions in your space – if people suddenly start asking new things (e.g., a new popular comparison or a new use case for your product), be the first to create content around it. Early movers can capture AI visibility before competitors even realise those queries are hot. Being comprehensive and quick to address new topics increases your odds of becoming the go-to source that AI answers rely on.
Optimise for Intent and Context, Not Just Keywords: Revisit your keyword strategy through the lens of user intent. With AI and advanced algorithms focusing on semantics, you want to make sure each piece of content fully satisfies the query’s intent. That means if you’re targeting “best running shoes for flat feet,” your content shouldn’t just repeat “running shoes for flat feet” a bunch of times. It should discuss what “flat feet” means, what features those users need, maybe expert quotes, reviews, comparisons – essentially, a one-stop authoritative answer.
Cover related questions (think “People also ask”) within your content to provide context. This broad, rich content tends to do well both in regular rankings and as source material for AI answers. Likewise, prune or revamp thin pages that only serve one tiny keyword purpose. Those might not cut it anymore. If an AI is choosing sources, it will likely pick the one or two best pages that cover a topic in depth. Aim to be that page. In practice, this could mean merging several small articles into one ultimate guide or adding sections to your product pages that address common pre-sale questions. It’s a more holistic approach to SEO content – think in terms of topics and entities, not just exact phrases.
Solidify Your Brand Authority: Build up your brand’s credibility both on and off your site, because authority signals will influence AI selection. This includes classic SEO off-page work: earning quality backlinks, mentions in reputable publications, and fostering positive reviews. LLMs trained on web data will “know” if your brand is frequently talked about in trustworthy contexts.
Google’s AI summaries also seem to favour sources that have authority on the topic (often mirroring E-E-A-T principles). As an eCommerce brand, make sure your site demonstrates expertise – have detailed About pages, author bios on content pieces, and links to any credentials or awards. These may or may not directly factor into an AI algorithm, but they certainly can’t hurt, and they build user trust. Additionally, monitor the narrative about your brand online. If there’s negative or incorrect information out there, it could surface in an AI answer (for example, an outdated forum post with a complaint might get picked up as “context”). Proactively address major criticisms or misconceptions through content on your site, so the correct story is out there for search engines to find. In short, invest in your brand’s digital footprint the way you’d invest in SEO – they are now deeply intertwined.
Prepare for AI as the “Front Door”: Finally, start thinking of SEO not just in terms of getting a click, but in terms of influencing the conversation that happens before the click. This is a mindset shift. Your content might be “consumed” partially or wholly on the search results page (via an AI summary or a featured snippet). So, ask yourself, if a user never visits us, what do we want them to know or feel about our brand?
Ensure that key selling points, unique differentiators, and calls to action are present in the content that could be pulled into an answer box. For example, if “free 2-day shipping” is one of your big advantages, weave that into product descriptions or FAQs – so if an AI summarises shipping options, your competitive edge isn’t missed. Treat those AI-generated blurbs as a first impression.
Additionally, look for ways to entice clicks when your content is featured. Some publishers are experimenting with hinting at further value – e.g., a snippet that teases a tool, calculator, or exclusive insight available on click. While you can’t control exactly what the AI will display, you can structure your content to naturally include “Learn more about X on our site” or have sections that prompt curiosity. The goal is to turn some of those zero-click situations back into clicks by offering something compelling beyond the basic answer.
No one has all the answers (not even the AI) on where search is heading, but the trends are clear. Search is evolving into a richer, AI-driven experience, and our SEO strategies must evolve with it. By focusing on technical excellence, content quality, and adaptive metrics, you can continue to grow your organic visibility and sales even as the ground shifts beneath us. The clicks you do earn in this new era may be fewer, but they will count more than ever. Now’s the time to get ahead of the curve – optimise for the world of AI and LLMs, and you won’t just protect your organic traffic, you’ll likely find new ways to expand your reach that your competitors are missing.
