Impacts of AI on Search Traffic

In 1999, the dotcoms were valued on traffic. IPO metrics revolved around eyeballs. Then Google launched AdWords, an ad model predicated on clicks, & built a $273b business in 2024. But that might all be about to change : Pew Research’s July 2025 studyreveals users click just 8% of search results with AI summaries, versus 15% without - a 47% reduction. Only 1% click through from within AI summaries. Cloudflare data shows AI platforms crawl content far more than they refer traffic back : Anthropic crawls 32,400 pages for every 1 referral, while traditional search engines scan content just a couple times per visitor sent. The expense of serving content to the AI crawlers may not be huge if it’s mostly text. The bigger point is AI systems disintermediate the user & publisher relationship. Users prefer aggregated AI answers over clicking through websites to find their answers. It’s logical that most websites should expect less traffic. How will your website & your business handle it? Sources: - Pew Research Center - Athena Chapekis, July 22, 2025 (https://lnkd.in/gKTqJ9iw) - Cloudflare: The crawl before the fall of referrals (https://lnkd.in/gqa26PUa) - Cloudflare Radar: AI Insights - Crawl to Refer Ratio (https://lnkd.in/gKP427sb) - Podcast: The Shifting Value of Content in the AI Age (https://lnkd.in/gUTkmPEz)

Google claims AI Overviews are revolutionizing search behavior, with Sundar Pichai boasting about "longer and more complex queries" and increased ecosystem traffic. The data tells a different story. 🔍My analysis of Similarweb data (based on 5+ billion search queries) reveals Google's claims are partially true but significantly oversimplified. Users visit Google US more frequently (+9%) but spend less time per visit - the digital equivalent of speed dating rather than deep engagement. The most revealing metric? Query length barely changed - growing from 3.27 to 3.37 words over two years in the US. That's a mere 3% increase, hardly the revolutionary shift Google describes. ⚖️Regional differences expose further contradictions. While US visits increased post-launch, UK metrics remained flat despite identical technology. AI Overviews are creating a "quick answer" behavior pattern rather than deeper search engagement. This matters because it challenges the narrative that AI Overviews represent a revolutionary enhancement. They're actually transforming interaction patterns in ways Google hasn't acknowledged. 💡For marketers and SEOs: The new reality is "visit more, stay less." Users are becoming increasingly selective about clicking through, making the quality of each site visit more critical than ever. 👉🏻Full analysis (charts, details, takeaways): https://lnkd.in/dX73DiTy

I've spent the last week doing (probably way too much) research for a client article on how AI is currently impacting SEO — and this is one of my biggest takeaways: LLMs and Google AI Overviews are leading to lower clickthrough rates and overall traffic, but especially for LLMs, the visitors who do end up on B2B sites tend to be much more likely to convert. Or, as a new study from Ross Hudgens at Siege Media (that inconveniently dropped right after I submitted my article) puts it: "LLMs are taking top funnel traffic from us, but seem to be sending us more bottom funnel (to the right places). This is a positive and should be weighed accordingly." I know there are plenty of legitimate qualms about how AI is scraping and using our painstakingly created content. But right now, if you're responsible for driving high-value leads to your website, optimizing your content for LLMs has to be part of your job. That's my other big takeaway from this research: the actual number of people using LLMs instead of traditional search isn't huge right now for most industries, but it's growing. Fast. Getting your brand to show up in LLMs certainly isn't going to matter LESS anytime soon. So, you might as well figure it out before all of your competitors do.

I hardly ever use Google search anymore, and I know I’m not alone. AI Search traffic is set to surge from 0.25% of search volume today to 10% by the end of 2025, and we’re already seeing inbound traffic from organic search crater for many websites. But beneath this headline lies a big shift in how potential customers will discover and evaluate products. When users interact with AI search tools like Perplexity or Gemini, they're not browsing multiple websites—they're getting synthesized answers in a conversational format. That means: - Your meticulously crafted SEO strategy becomes largely irrelevant in AI search - The context and accuracy of how AI agents and LLMs understand your product determines your visibility - New tools are required for understanding and optimizing how your content is showing up in AI search tools. Using the model APIs alone doesn’t work because their data is static. For developer-facing companies, the stakes are even higher. When developers use GitHub Copilot, Cursor, Replit, or other AI coding agents, these tools need to understand your API documentation with accuracy. If these tools can’t properly index and parse your developer docs, there are severe consequences: agents will incorrectly implement your APIs, developers waste hours debugging non-working code, and ecosystem adoption slows. Two interesting companies that have emerged to help solve this: 1) Profound is building essential intelligence infrastructure that reveals how AI search tools present your brand through their Answer Engine Insights. Their tools help to measure how your company appears in live AI search products like ChatGPT, and provide insights on how to improve and optimize your content’s visibility. 2) Mintlify creates exceptional developer docs that are also optimized for AI coding agents through automatic generation of llms.txt files. This is an emerging standard to help optimize web content for consumption and retrieval by LLMs, and they’re helping to drive the adoption by making this shameless for their customers. This ensures their customers like Anthropic, Replit, and Perplexity have API documentation that AI coding assistants can accurately parse and use. For any founders or marketers not thinking about this and implementing llms.txt and llms-full.txt files, now is the time. Are there other tools or tactics you’ve found that are helping to monitor and improve how they appear in AI search results and conversations?

We just ran the largest AEO study for B2B brands. 117,432 leads analyzed. October 2024 to April 2025. One core question: Can AI search actually drive pipeline, not just clicks and visibility? Here’s what we found: - ChatGPT leads convert nearly 2x better than Google or Bing. It only drove 7.8% of leads, but delivered the highest lead-to-close rate across all channels. - Google still owns the top of the funnel (58% traffic share), but its downstream efficiency is slipping. - Perplexity performed well on quality, but at a fraction of ChatGPT’s volume. ChatGPT gets 6x more traffic. - Gemini ranked lowest in traffic and lead contribution among AI search platforms. Why this matters: - LLMs are rapidly becoming the new gatekeepers of information on the web, the intermediary between brands and buyers. - The funnel is collapsing, buyers are discovering products and making decisions directly inside ChatGPT and Perplexity. - Leads from LLMs trust the answers more and it shows. They convert better at every stage of the funnel. - If you’re not monitoring how AI search is impacting your brand and funnel, you’re ignoring the most valuable source of leads on the web today. Full study is linked in the comments. Curious to hear what patterns you’re seeing in your funnel- drop a comment below!

At Duda, we’ve analyzed traffic across more than 1 million websites published on our platform, and the trend is clear: AI-driven traffic is growing extremely fast, especially from ChatGPT.  That doesn’t mean SEO is dead. Far from it, Google is still generating significantly more traffic to websites than all AI platforms combined. But it does mean that web professionals need to adapt their strategies to make sure their clients’ websites are discoverable in both traditional *and* AI search. The good news is a lot of the best practices to rank for traditional search are still relevant in the age of AI discoverability - with some new considerations… ✅ What still works: Authoritative, well-structured content, following E-E-A-T principles   Strong technical SEO (Core Web Vitals, mobile responsiveness, fast load times) Focus on genuinely helpful content, not just ranking tricks 🆕 Where to renew focus in the age of AI: LLMS.txt files (think robots.txt, but for AI) to help LLMs understand your site Structured data like Local Business, FAQ, and Product schema to surface in AI answers IndexNow & Google Search Console integrations to get new content discovered quickly Server-side rendering (SSR) so content isn’t hidden behind JavaScript 🚫 What to avoid: Keyword stuffing Generic backlink strategies Prioritizing visibility over value At Duda, we’re building tools to help agencies and SaaS platforms stay ahead of the curve, including being the first website builder to automatically generate LLMS.txt for all sites built on our platform. If you’re a web pro building websites at scale, it’s time to optimize for both search engines and AI assistants. Learn more about how to future-proof your sites for AI discoverability here: https://lnkd.in/dA_3YrG6

When I started in SEO, traffic was everything. Then, SEO changed. Google got scary good at predicting what content would satisfy a search. So I started optimizing for the searcher, not the search engine. Today, your customers are using different search engines for different types of searches. For informational searches: generative AI platforms like ChatGPT and Perplexity deliver a better experience -- At best, these platforms feel like a natural language conversation with a knowledgeable friend or trusted expert. For transactional searches: traditional search engines like Google work better, helping you find product pages, local businesses, and specific websites. Today AI search platforms send less traffic, but still represent a huge opportunity to capture organic demand. It’s time to change our SEO and content strategies for gen AI search. Here are 5 changes I would make now: ❌ Stop ignoring zero-click searches because they don’t appear in analytics ✅ Embrace zero-click searches as opportunities to build brand awareness ❌ Stop worrying about rankings and traffic volume ✅ Get forensic about what converts traffic into customers — then  double-down on THAT ❌ Stop publishing tons of informational long-form blogs ✅ Use AI to create interactive content that helps guide customers to a purchase decision ❌ Stop overspending on performance marketing ✅ Overspend on audience research, community engagement and brand building ❌ Stop relying on keyword-research tools to create generic content ✅ Talk to your customers to better understand the questions they ask when making a purchase decision What do all these changes have in common? Focus MORE on customers and LESS on search engines. The key is to understand YOUR customers’ search behavior. Do your research: Start by asking your customers how they search for information about your products. Stop obsessing over traffic and start targeting the search intents that drive conversions. More information: Download our SEO for Revenue Guide (link in the comments).

📉 AI Overviews are killing clicks — new study confirms it. Last week, Ahrefs just dropped a bombshell analysis: After reviewing 300,000 informational keywords, they found that the introduction of Google’s AI Overviews led to a 34.5% drop in clickthrough rates (CTR) for top-ranking pages. Here’s what they found: • Average CTR for position #1 on informational queries dropped from 5.6% to 2.6% when AI Overviews were present • That’s a 34.5% decrease in clicks for top-ranking pages • Even when pages are cited in the Overview, they’re competing with multiple other links, reducing the chance of earning traffic Despite Google’s claim that AI Overviews increase clicks when links are included, there’s still no way to isolate these metrics in Search Console. We’re flying blind. This shift feels very similar to what happened with Featured Snippets — except broader, more opaque, and more disruptive. If your content strategy still treats position #1 as a guaranteed click magnet, it may be time to rethink how you measure success. If your business depends on organic search, now is the time to reassess your visibility, rethink attribution, and diversify your traffic sources. SEO isn’t dead — but the SERP is getting a lot more crowded (and a lot less clickable). Kudos to Ryan Law and Xibeijia Guan at Ahrefs for digging into this. Worth a read for anyone watching the SERP evolve. Read the full Ahrefs study here: link in comments

some say AI will make brand irrelevant. the latest data suggests the exact opposite. while AI Overviews are crushing generic search, they’re actually boosting branded search. the gap between them is widening, fast. first, the bad news for generic, non-branded terms: -> AI Overviews now appeared on 13% of U.S. searches in march (up 2x since january). -> when they show up, click-through rates (CTR) for non-branded keywords fall nearly 20%. -> for some top publishers, traffic is already down ~40%. now, the good news for brands: -> branded searches rarely trigger an AI Overview (only ~5% of the time). -> and when they do, branded CTR increases by almost 19%. the takeaway is simple. the value of ranking for your category is collapsing. the value of being searched for by name is skyrocketing. this changes where we must invest our time and energy. away from the shrinking returns of generic SEO and toward building a brand that people seek out directly. here's a quick check for your own brand: -> are your branded search impressions growing faster than your YoY revenue? -> are your branded search clicks (paid + organic) growing faster than your impressions? our job is to own a word in our customers' minds and convert that mental real estate to baseline revenue. that's the moat in the AI era.

➡️ While you’re still chasing page-one rankings, ChatGPT & friends are quietly siphoning off your traffic—and your leads. The world is moving away from "Search Engines" to "Answer Engines" ➡️ Here are the 10 SEO shockwaves you need to address **Traffic Diversion from Organic Results to AI Summaries**  Search engines now prioritize AI-generated answers (e.g., Google’s AI Overviews) over traditional website links. **Lower Engagement from LLM-Generated Traffic**  Users arriving via LLM referrals show **significantly lower engagement** than organic search visitors. . **Attribution and Tracking Complexity**  LLMs obscure traditional referral paths, making it harder to track user journeys, measure conversions, or attribute leads accurately. T 4. **Shift from SEO to Generative Engine Optimization (GEO)**  Traditional SEO (focused on keywords/backlinks) is becoming obsolete. GEO prioritizes **authoritative content** optimized for LLM citation 5. **Increased Costs for Multi-Platform Optimization**  Brands must now optimize for both traditional search engines **and** LLMs (e.g., ChatGPT, Perplexity, Claude). This requires dual-strategy investments in SEO and GEO, straining budgets. **Commoditization of Top-Funnel Content**  Informational content (e.g., "What is X?") is rarely cited by LLMs, as answers are synthesized without attribution. This devalues top-funnel content, forcing brands to focus on bottom-funnel queries (e.g., "best tools for X") where LLMs explicitly recommend products . **Erosion of Backlink Value**  Backlinks remain vital for traditional SEO but matter less for LLM visibility. LLMs prioritize **semantic relevance** and comprehensive topic coverage over domain authority. **Local SEO Fragmentation**  Local businesses face reduced visibility as LLMs struggle to integrate real-time local data. . **Keyword Intent Misalignment**  LLMs interpret **conversational queries** contextually, rendering traditional keyword research inadequate. **Brand Authority Dilution**  LLMs rarely display bylines or source brands prominently, weakening brand-driven trust signals. ➡️ The transition demands a hybrid approach: **traditional SEO for residual organic traffic** and **GEO for LLM visibility**. ➡️ Join us TONIGHT at 7p ET / 4p PT at our weekly Learning Lab (a video call) with a discussion this topic with Rosemary Brisco, AI Training Guru and Mike Ensing. Sign up via Meetup David Edelman Bill Hobbib Kim Pitsko John Sviokla Liz Vanzura John Spottiswood Parag Shah Dan Slavin #SEObloodbath #GEO #AEO