# ChatGPT vs Google: how people actually split their searches in 2026 **URL:** https://organikpi.com/blog/geo-ai-search/chatgpt-vs-google-search-behavior/ **Published:** 2026-07-30 **Modified:** 2026-07-30 **Author:** Daniel Shashko > Google handles more than 5 trillion searches per year, while ChatGPT reached 700 million weekly active users in September 2025. Datos and SparkToro measured all AI tools at 3.2% of US desktop searches in Q4 2025, with Google alone at 73.7%. People move research, advice, and writing tasks to ChatGPT while navigation, shopping, and quick lookups stay on Google. Brands need visibility in both channels, tracked separately. --- > Google handles more than 5 trillion searches per year, while ChatGPT reached 700 million weekly active users in September 2025. Datos and SparkToro measured all AI tools at 3.2% of US desktop searches in Q4 2025, with Google alone at 73.7%. People move research, advice, and writing tasks to ChatGPT while navigation, shopping, and quick lookups stay on Google. Brands need visibility in both channels, tracked separately. People use both, and they split the work by task. Google reports [more than 5 trillion searches per year](https://blog.google/products/ads-commerce/google-search-ai-brand-discovery/), while all AI tools combined took [3.2% of US desktop searches](https://sparktoro.com/blog/new-research-search-happens-everywhere-an-analysis-of-41-websites-with-significant-search-activity/) in Q4 2025. ChatGPT absorbs research, advice, and writing; Google keeps navigation, shopping, and quick lookups. **The ChatGPT vs Google split comes down to task type.** Google Search remains the default for finding websites, products, and places, at a scale above 5 trillion queries per year. ChatGPT works as an advisor that people ask for guidance, research help, and writing. Brands now need measured visibility in both. ## The headline numbers, side by side Google said in May 2025 that it sees more than 5 trillion searches annually, a figure from its own internal data. That works out to billions of searches every day. No other platform on the web operates anywhere near this volume. Google also reported that AI Overviews drive over 10% more Google usage for the query types where they appear, based on internal tests in the US and India. By Google’s own measurement, AI answers are growing Google usage. OpenAI said in September 2025 that ChatGPT reached [700 million weekly active users](https://openai.com/index/how-people-are-using-chatgpt/). Engagement deepens over time as well: [OpenAI Signals data](https://openai.com/index/how-chatgpt-adoption-has-expanded/) published in June 2026 shows users send 50% more messages per day six months after signing up than they did when they signed up. The growth is also going global. The same Signals release shows users who mostly write in a language other than English now make up over half of active ChatGPT users, and the earlier study found adoption growth rates in the lowest income countries running over 4x those in the highest income countries by May 2025. ChatGPT is widening its base while Google defends an installed habit. The two headline numbers measure different things. Weekly users describe reach. Search volume describes frequency. Any honest comparison keeps those labels attached. ## ChatGPT’s real share of search Datos and SparkToro measured search behavior across 41 major websites, using a desktop panel of millions of US devices. [Google alone accounted for 73.7% of all desktop searches](https://sparktoro.com/blog/new-research-search-happens-everywhere-an-analysis-of-41-websites-with-significant-search-activity/) across those sites in Q4 2025. Every AI tool combined held 3.2%. The same study found Amazon, Bing, and YouTube each receive more desktop search activity than ChatGPT. Rand Fishkin’s conclusion: Google’s real US share sits closer to 70% than the 90% you see in headline trackers, once you count every place people actually run queries. The category split makes the point sharper. About 80% of desktop searches still happen on traditional search engines. Commerce sites like Amazon and Walmart take roughly 10%. Social networks hold about 5.5%. AI tools, all of them together, sit at 3.2%. In the EU and UK, Google is even more dominant, and ChatGPT runs slightly higher than its US share. Desktop search share by category, Q4 2025. Data: Datos x SparkToro panel across 41 sites. Google did shrink. It lost 3.5 points of desktop search share over 2025 in that panel, spread across commerce sites, Bing, YouTube, and AI tools. Movement is real; collapse is nowhere in the data. ChatGPT’s slice is a working channel and a small share of the total at the same time. We keep a running picture of these shifts in our [quarterly look at where AI search actually stands](https://organikpi.com/blog/geo-ai-search/state-of-ai-search/). For the other fast-growing engine, see [how Perplexity stacks up against Google](https://organikpi.com/blog/seo-strategy/perplexity-vs-google-market-share/) on the same kind of panel data. ## Which tasks move to ChatGPT, and which stay on Google OpenAI answered this with its own data. Working with Harvard economist David Deming, the company analyzed 1.5 million ChatGPT conversations. [Three quarters of conversations](https://openai.com/index/how-people-are-using-chatgpt/) focus on practical guidance, seeking information, and writing. ### The three buckets in OpenAI’s study - Asking, 49% of messages: advice, recommendations, and judgment calls. This bucket behaves most like search. - Doing, 40%: drafting text, planning, programming, and other task completion. - Expressing, 11%: reflection and play, with no search equivalent at all. How ChatGPT messages split by task. Data: OpenAI x David Deming, 1.5 million conversations. Only about 30% of consumer usage is work related; roughly 70% is personal. The average ChatGPT session looks like life admin, study help, and writing support. What stays on Google: navigation, local searches, and most transactions. Nobody prompts a chatbot to reach their bank login, and shopping queries still flow to Google and Amazon at far larger volume. Quick single-fact lookups mostly stay too, because the address bar is faster than a chat. Google’s own results show the same pattern from the other side. Pew found that only 8% of one- or two-word searches triggered an AI summary, while 53% of searches with ten or more words did, and 60% of searches that started with question words produced one. Long, conversational queries get AI answers wherever people type them. The queries that move are the messy ones: multi-constraint research, comparisons, and first drafts that used to take a whole Google session. That shift is why [prompt research works differently from keyword research](https://organikpi.com/blog/seo-strategy/prompt-research-vs-keyword-research/), and why we spend real time on [getting brands recommended inside ChatGPT answers](https://organikpi.com/blog/geo-ai-search/how-to-rank-in-chatgpt-search/). ## ChatGPT vs Google at a glance The table sticks to primary-source figures and labels what each one measures. DimensionChatGPTGoogle SearchScale700 million weekly active users (OpenAI, September 2025)More than 5 trillion searches per year (Google, 2025)Dominant tasksPractical guidance, information seeking, writing; 49% of messages are advice-style askingNavigation, local, shopping, quick lookups, plus high-volume everything elseCitation behaviorLinks appear when it runs a live search; memory-based answers often show noneAI Overviews appear on about 18% of searches; 1% of those visits click a cited sourceTraffic sent to websitesSmall referral stream that needs its own tracking setupDominant referrer for most sites; clicks drop from 15% to 8% when an AI summary appearsAd modelNo ad marketplace at Google’s scale yetFull ad auction; ads run in AI Overviews and are being tested in AI Mode ## Citations and trust Google’s AI Overviews cite sources by design. [Pew’s browsing study](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/) found Wikipedia, YouTube, and Reddit are the most cited domains, together making up 15% of AI summary sources. Government sites appear in 6% of AI summary citations, versus 2% of standard results. The typical AI summary ran 67 words, and 88% of them cited three or more sources. ChatGPT links out when it runs a live web search. Answers written from model memory carry no citations, and that is where wrong facts slip through with nothing to correct them. The defense is boring and effective: publish clear, current pages with named sources, and build [strong author signals](https://organikpi.com/blog/brand-authority/eeat-ai-search-author-authority/) that engines can verify. ## What happens to clicks and referral traffic Pew Research Center tracked the real browsing of 900 US adults in March 2025. Users clicked a traditional result on [8% of searches that showed an AI summary, versus 15% without one](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/). Clicks on sources cited inside the summary happened on just 1% of those visits. People were also more likely to end their browsing session entirely after seeing an AI summary: 26% of those visits, against 16% on standard result pages. Answers on the results page keep people on the results page. We broke down the traffic consequences in our analysis of [what AI Overviews do to organic click-through](https://organikpi.com/blog/seo-strategy/ai-overviews-ctr-impact/). The short version: informational queries lose the most clicks, while commercial queries hold up better. ChatGPT sends its own referral stream, and default analytics handles it badly. Set up [proper tracking for AI assistant referrals in GA4](https://organikpi.com/blog/technical-seo/ga4-ai-search-referral-attribution/) before judging the channel. And since clicks shrink either way, build [a plan for being seen without the visit](https://organikpi.com/blog/geo-ai-search/zero-click-ai-search-strategy/). ## What search share numbers actually measure Most viral charts on this topic compare things that don’t match. Traffic to chatgpt.com counts visits. Search share counts queries. The Datos panel found only about half of ChatGPT visitors enter a prompt at all; many arrive just to read a chat someone shared with them. Popularity charts built on visits will mislead you about usage. Counting choices matter too. That study treats a whole multi-prompt conversation as one search, the same way a Google search with ten clicks counts once. The panel is desktop only, and a huge share of search happens on mobile. Every share number you read carries choices like these. One more twist: most AI answers happen on Google. AI summaries appeared on 18% of the Google searches Pew studied, and at Google’s volume that makes Google the largest AI answer engine by an order of magnitude, a point Rand Fishkin makes from the Datos data. Meanwhile the Datos panel shows under 0.1% of searchers clicking over to AI Mode. For most brands, [how AI Overviews and AI Mode differ](https://organikpi.com/blog/geo-ai-search/ai-overviews-vs-ai-mode/) matters more than the ChatGPT share debate. ## What this means for your channel mix Treat them as two channels with different jobs. Google drives the visible volume and the attributable revenue. ChatGPT shapes the research phase you mostly can’t see, especially for B2B shortlists and considered purchases. That split is why we run classic SEO and [generative engine optimization](https://organikpi.com/blog/geo-ai-search/what-is-geo-generative-engine-optimization/) as separate tracks with shared foundations. The same page rarely wins both by accident. Our guide to [how much budget belongs on each side](https://organikpi.com/blog/geo-ai-search/geo-vs-seo-budget-split/) covers the tradeoffs by business type. The ad picture is splitting too. Google confirmed in May 2025 that ads in AI Overviews were expanding to desktop in the US, and that it had started testing [ads inside AI Mode responses](https://organikpi.com/blog/geo-ai-search/ads-in-google-ai-mode/). ChatGPT has nothing comparable at that scale, which keeps organic visibility as the main way into its answers for now. Measurement changes with the mix. Rankings say nothing about whether ChatGPT recommends you. Track [your brand’s presence across AI answers](https://organikpi.com/blog/seo-strategy/ai-search-brand-share-of-voice/) alongside rankings, and re-test after major model releases, since [answer patterns shifted noticeably with GPT-5](https://organikpi.com/blog/geo-ai-search/gpt5-chatgpt5-seo-content-impact/). ## Where to start this week - Pull GA4 referrals and isolate chatgpt.com, perplexity.ai, and gemini.google.com. Small numbers still set your baseline. - Ask ChatGPT, AI Mode, and Perplexity ten buying questions from your category. Note which brands get named and which pages get cited. [Running this audit yourself](https://organikpi.com/blog/geo-ai-search/diy-ai-brand-visibility-audit/) takes about an hour. - Rewrite your highest-value pages to lead with direct answers, atomic stats, and named sources. That format survives both engines. - Add structured data where it earns its keep. Our notes on [which schema types correlate with AI citations](https://organikpi.com/blog/technical-seo/faq-howto-article-schema-ai-citations/) cover the short list worth implementing. If you would rather not run those prompts by hand, [our free GEO/AEO Tracker scores your brand](https://organikpi.com/tools/geo-aeo-tracker/) across ChatGPT, AI Overviews, Perplexity, and three more engines, with the exact prompts and citations behind every score. Google remains the bigger channel in 2026. The brands winning both treat the split as a planning decision, and they start now. ## Frequently Asked Questions ### Is ChatGPT replacing Google in 2026? No. Datos and SparkToro measured Google at 73.7% of US desktop searches in Q4 2025, with all AI tools combined at 3.2%. Google lost 3.5 points of share during 2025, and that loss spread across Amazon, Bing, YouTube, and AI tools. ChatGPT changes which tasks people search for, and Google still carries the volume. ### How many people use ChatGPT compared to Google? OpenAI reported 700 million weekly active ChatGPT users in September 2025. Google reports more than 5 trillion searches per year, which works out to billions of searches every day. The two figures measure different things, reach versus frequency, so treat direct comparisons between them carefully. ### What do people use ChatGPT for instead of Google? OpenAI's study of 1.5 million conversations found three quarters focus on practical guidance, seeking information, and writing. About 49% of messages are advice-style questions. Multi-step research, comparisons, and first drafts move to ChatGPT, while navigation, local searches, and most shopping queries stay on Google or Amazon. ### Does ChatGPT send traffic to websites? Yes, though far less than Google. ChatGPT cites and links sources when it runs a live search, and many answers come from model memory with no links at all. Referrals show up in GA4 under chatgpt.com, and default settings often bucket them poorly, so set up dedicated tracking before judging the channel. ### Do Google's AI Overviews reduce clicks to websites? Yes. Pew Research Center found users clicked a traditional result on 8% of searches that showed an AI summary, versus 15% without one. Only 1% of visits produced a click on a source cited inside the summary. Informational queries feel the drop hardest, which pushes brands toward zero-click visibility. ### Should I optimize for ChatGPT or Google first? Start with Google if you rely on traffic and revenue today, since it still carries most search volume. Add ChatGPT work once your category shows AI-assisted research, which covers most B2B and considered purchases in 2026. The overlap is real: clear structure, cited sources, and direct answers help in both places.