ChatGPT Sends 21% of Its Traffic to Google. Here’s Why That Matters.

Are you still treating ChatGPT as a novelty — something your team experiments with but your marketing strategy ignores?

A new study from Semrush — analysing over a billion lines of clickstream data across 17 months — has just delivered the clearest picture yet of how ChatGPT is reshaping the search journey. And the headline finding is one that should make every marketing leader sit up: more than one in five of all ChatGPT referral clicks go directly to Google.

That’s not a quirk. It’s a pattern — and it’s growing.

The share of ChatGPT’s outbound traffic going to Google climbed from roughly 14% at the start of the study to over 21% by early 2026.

For marketers who’ve been watching the AI search landscape unfold and wondering where ChatGPT fits in, this data answers that question decisively. ChatGPT is now a core part of the buyer journey — not replacing Google, but feeding into it, reshaping how people research, compare, and ultimately choose who to buy from.

Charlie Marchant, Exposure Ninja’s CEO, puts it plainly: “ChatGPT is a new search channel. People are searching within ChatGPT. It’s now part of the search journey to finally find the business or the product that they actually want.”

In this article, we’ll break down what the Semrush data actually shows, what it means for your AI Search Optimisation strategy, and the practical steps you should be taking to make sure your brand is visible where your buyers are now spending their time.

Executive Summary

Semrush’s study tracked over a billion lines of US clickstream data from October 2024 to February 2026, covering a 200-million-user panel. The findings confirm that ChatGPT has become a meaningful and growing source of referral traffic — but also that it functions very differently from traditional search.

• Referral traffic from ChatGPT to websites grew 206% year on year (January 2025 vs January 2026), even as ChatGPT’s own traffic plateaued near a billion monthly visits.
• Over 30% of that referral traffic goes to just 10 domains, with Google alone accounting for 21.6%.
• Only 34.5% of ChatGPT queries now trigger a live web search — down from 46% in late 2024 — meaning most responses still rely on training data.
• And between 65% and 85% of ChatGPT prompts don’t match any traditional search keyword at all.

For marketing leaders, the implication is clear: ChatGPT is absorbing more and more of the research and consideration phase.

Users are making purchase decisions inside their AI conversations before they ever click through to a website.

If your brand isn’t visible in those conversations — both in ChatGPT’s training data and in the sources it pulls from when it does search the web — you’re invisible at the moment that matters most.

What Does the New Semrush ChatGPT Data Actually Show?

The scale of this study is worth emphasising. Semrush analysed more than a billion lines of clickstream data across both mobile and desktop devices, drawn from a panel of 200 million users. That’s not a survey or a sample of ChatGPT conversations — it’s a detailed map of where people go before, during, and after interacting with ChatGPT, tracked across 17 months.

The headline numbers are striking. ChatGPT referral traffic — the visits it sends to other websites — grew 206% comparing January 2025 to January 2026. The number of unique domains receiving referral traffic from ChatGPT expanded from around 71,000 per month in October 2024 to a peak of roughly 260,000 in October 2025, before settling at approximately 170,000 by February 2026.

As Charlie explains: “Lots of people have been talking about, ‘Oh my God, ChatGPT doesn’t actually send traffic anywhere. Is it just a waste of time, marketing for ChatGPT?’ It’s one of the biggest objections I hear to actually creating a search strategy to optimise for ChatGPT. And finally, we have some data on a large scale that shows, oh my goodness, yes — ChatGPT is referring out traffic.”

Meanwhile, ChatGPT’s own traffic has plateaued. After explosive growth through most of 2025, total visits levelled off around November 2025, hovering near a billion monthly visits through early 2026. But even as fewer new users arrive, those who remain are using it more intensively — average queries per session jumped 50% in the final four months of the study period.

This is a platform that’s maturing, not declining. The casual experimenters may have moved on, but the committed users — your potential customers — are going deeper.

Why Is ChatGPT Sending So Much Traffic to Google?

The 21% figure raises an obvious question: if people are using ChatGPT, why are they ending up on Google?

The answer lies in how people actually use these tools together. ChatGPT is absorbing the research phase — the broad, exploratory, “help me understand my options” part of the journey. But when users are ready to make a purchase decision, compare specific providers, or navigate to a particular website, they’re turning to Google to complete that final step.

Charlie sees this pattern in Exposure Ninja’s client data across more than 60 industries: “Searchers who use Google move between the platforms. When people are researching top level, they start in ChatGPT because they can ask something huge, they can write a paragraph in there. But Google’s the platform we trust. When we’re actually ready to make a purchase decision or navigate to a website to compare two different providers, we’re gonna go to Google.”

This isn’t a sign that ChatGPT is failing. It’s a sign that the search journey is now multi-platform — and ChatGPT is a massive influencer in the decision-making process, even when the final click happens elsewhere. Exposure Ninja’s own AI search research found that more than 80% of people are making at least half of their purchase decisions during their AI conversations, before they ever visit a brand’s website.

For your SEO strategy, this means something important: even if the click comes from Google, ChatGPT may have been the tool that shaped which brand the user searched for. If you’re only optimising for Google, you’re optimising for the end of the journey — and missing the platform where the real decision is being made.

Does ChatGPT Actually Send Traffic to Websites?

Yes. Decisively yes. And this is where the Semrush data is most useful, because it puts hard numbers behind something that’s been hotly debated.

The 206% year-on-year increase in referral traffic isn’t a marginal gain. It represents a fundamental shift in ChatGPT’s role — from a tool people use to get answers, to a platform that actively drives users toward websites, products, and services.

But there’s a nuance worth understanding. Only 34.5% of ChatGPT queries trigger a live web search as of February 2026, down from 46% in late 2024. That means the majority of ChatGPT responses are still drawing on its training data — the knowledge baked into the model — rather than searching the web in real time.

Charlie explains what this means in practice: “For a lot of prompts, ChatGPT still doesn’t use live search. It’s only around 35% of queries. That means most of the ChatGPT responses really are still relying on the training data, the model knowledge, rather than live web retrieval. That means that as that training data gets updated, it does take a bit more time if you’ve been optimising for ChatGPT to show in those types of answers.”

This has practical implications for your optimisation timeline. When ChatGPT does search the web, the benefits of your AI Search Optimisation strategy can show up relatively quickly. But for the majority of queries — the ones answered from training data — visibility depends on building a strong enough digital footprint that your brand is captured in the next training data update.

This is why a comprehensive approach matters. Your own website content, your third-party media coverage, your reviews, your presence on platforms like Reddit and YouTube — all of these feed into the model’s understanding of your brand. It’s not enough to optimise one channel and hope for the best.

Where Does ChatGPT Referral Traffic Go?

The distribution of ChatGPT’s referral traffic tells us a lot about how the platform is being used — and where the opportunity lies.

Thirty percent of all referral traffic goes to just 10 domains. Google dominates at 21.6%, followed by YouTube, GitHub, Amazon, Facebook, LinkedIn, DeepSeek, and the National Institute of Health. These are massive platforms with existing audiences — not niche websites.

But that still leaves 70% of referral traffic distributed across a vast number of other websites. And it’s in this long tail where most businesses will find their opportunity.

Charlie uses a practical example: “If I’m looking for the best CRM software for a small to medium-sized enterprise, I want to actually be comparing different providers inside ChatGPT. I want to be looking at HubSpot, Salesforce, all of the alternatives. And if you are one of those businesses, you absolutely want your business to be in that list, and you absolutely want to be getting that relevant referral traffic to your website.”

The industries receiving the most ChatGPT referral traffic align with its user base: online services, computer software and development, mass media, and — notably — education. ChatGPT’s audience skews younger and more tech-savvy, which is worth bearing in mind when assessing whether this channel is relevant for your specific customer base.

The critical action here is to check your own data. Tracking your brand’s visibility in LLMs will show you exactly how much referral traffic and how many conversions are coming from ChatGPT versus other AI platforms — and the answer might surprise you.

How Are ChatGPT Prompts Different From Google Searches?

One of the most significant findings in the Semrush study is just how different ChatGPT prompts are from traditional search queries — and what that means for content strategy.

Between 65% and 85% of ChatGPT prompts couldn’t be matched to any traditional search keyword in Semrush’s database of over 27 billion keywords. These aren’t keyword searches. They’re conversations.

Where a Google search might be “best CRM software,” a ChatGPT prompt might be “I have a small pharmaceutical company and I need a CRM that handles compliance requirements and integrates with our existing ERP system — what are my best options?” The intent is the same. The language is completely different.

Charlie explains why this is actually good news for businesses with clear positioning: “This is perfect. This is why ChatGPT’s traffic is often considered to be much higher intent. People can be more specific. If you as a business are really clear on your customer personas, who buys from you, who you’re targeting, who your most profitable customers are, then you are much more likely to show up for those specific types of customers.”

The Semrush data adds another layer to this. Among the ChatGPT prompts that do match traditional search terms, most have navigational or transactional intent — people who already know what they want and are using ChatGPT to get there. The informational and commercial research queries, meanwhile, are happening in a completely new language that traditional keyword tools don’t capture.

This is why topic research for AI platforms requires a different approach. Instead of targeting specific keywords, you’re thinking about the topical areas you want to own — the types of questions your customers are likely to ask, the language they use, and the context they provide when they’re exploring options.

Charlie puts it succinctly: “We’re not just thinking about the keywords we want to show up for on Google. We’re creating a really good topic cluster architecture around the topical areas we really want to own, the types of prompts that we want to show up for.”

Should You Still Optimise for Google?

Absolutely. And the Semrush data makes the case more clearly than ever.

If 21% of ChatGPT’s referral traffic goes to Google, then your Google presence is still the foundation that everything else builds on. When someone researches options in ChatGPT and then goes to Google to compare providers or navigate to a specific website, you need to be there.

But the way you think about Google performance needs to evolve. The era of chasing traffic volume as a primary metric is over.

Charlie is blunt about this: “What I think is over is the way of thinking about ‘if we produce way more content and we get way more sessions, then if we keep our conversion rate at roughly 2%, we’re gonna get way more leads.’ Actually, we are gonna see a lot of that top-of-funnel traffic not coming to the website. Organic traffic — a lot of people are gonna be seeing, even though impressions are up, clicks are actually down.”

The shift is from measuring clicks to measuring conversions. If your organic traffic is flat or even declining slightly, but your conversion rate is improving because the visitors who do arrive are further along in their decision-making process (thanks to ChatGPT doing the research for them), that’s a win.

Being visible in AI Overviews is equally important. As Google continues rolling out AI-powered search features, the businesses that appear in those AI-generated answers at the top of the results page will capture the lion’s share of attention — and clicks.

Your SEO foundation isn’t becoming less important. It’s becoming the prerequisite for everything else.

How Should You Optimise for ChatGPT Differently?

Optimising for ChatGPT isn’t a separate discipline from SEO — but it does require a broader lens and some specific adjustments.

The first priority is your own website content. AI platforms, including ChatGPT, need to understand what your business does, who it serves, and why it’s the best option. If you’re an e-commerce business, this means having well-defined product categories, strong category overview pages, and clear product pages. If you’re a service business, it means dedicated landing pages for each target demographic and market.

As Charlie explains: “AI can understand if you are a global player. They can understand if you are in lots of different markets. They can understand if you have a product stack relevant to multiple different sectors. But they won’t know that if you don’t have content on your site that clearly indicates what those markets are, what those sectors are, or what those products actually are and who they’re for.”

The second priority is your third-party presence. ChatGPT doesn’t just look at your website — it considers the entire digital ecosystem around your brand. Your media coverage, your reviews on platforms like Trustpilot and G2, your presence on Reddit and YouTube, your mentions in industry publications — all of this feeds into whether ChatGPT recommends you or not.

Running an AI Search Optimisation audit will help you identify exactly where your gaps are — both on your own site and across the third-party sources that AI platforms trust.

The third priority is understanding query fanout. When a user asks ChatGPT a complex question, the platform doesn’t just run a single search. It breaks the query into multiple sub-queries to build a comprehensive answer. Your content strategy needs to account for the underlying questions that sit beneath the surface-level prompt — and make sure you’re providing useful, authoritative answers to as many of them as possible.

For a detailed walkthrough of the strategies that are working right now, the best AI Search Optimisation strategies for 2026 covers the full framework we use with our clients.

Does a ChatGPT Strategy Work for Other AI Platforms?

There’s significant overlap — but the answer isn’t a simple yes.

Charlie is clear that the majority of work you do to optimise for ChatGPT will benefit your visibility across other AI platforms too. Strong website content, authoritative third-party mentions, clear positioning — these are the foundations that every AI platform relies on when deciding which brands to recommend.

But there are important nuances. Each platform has different preferences for which sources it trusts, how it weights authority, and which types of content it favours. A strategy that performs well in ChatGPT may not translate identically to Claude, Gemini, or Copilot.

Charlie explains: “There are very, very stark nuances in where these AI platforms prefer to source information, what websites they trust and feel are authoritative. A very strong search strategy would be taking that into account, thinking about those third-party sources that are trusted, and building a strategy around making sure you’re visible in the right ones.”

For B2B businesses in particular, the platform mix matters. Enterprise companies mandated to use Microsoft tools will default to Copilot. Tech-forward businesses may gravitate toward Claude. Consumer audiences are most likely to be on ChatGPT or using Google’s AI features.

The smartest approach? Check your Google Analytics data. You can split out referral traffic by AI platform to see exactly which ones are driving sessions and conversions for your specific business. That data should guide where you invest your optimisation effort — not assumptions about what the general market is doing.

If you’re building an enterprise AI strategy, platform-specific data is the difference between an efficient plan and an expensive guess.

Has ChatGPT’s Growth Peaked?

ChatGPT’s traffic has plateaued, but that doesn’t mean it’s peaked in terms of influence or marketing value.

The Semrush data shows total visits levelling off around a billion monthly visits from November 2025 onwards. At the same time, users are engaging more deeply — asking more prompts per session and having longer, more detailed conversations. The platform is maturing from mass-market novelty into a daily-use tool for its core audience.

Charlie attributes some of the plateau to increased competition: “The new models of Gemini are much more competitive and people have enjoyed using Gemini more than previously. And Claude’s position in the market as well — Claude have very much gone after the enterprise, the workplace users, the developers.”

But she’s equally clear that this isn’t the end of the road: “They’re still huge. There’s quite a lot of ground that Claude or Gemini or any of the other AI chatbots would actually have to regain to get up to ChatGPT’s market share.”

Google, of course, holds the strongest competitive position — owning both the world’s dominant search engine and Gemini. The interplay between these platforms will continue to evolve. But for now, ChatGPT remains the largest AI chatbot by a significant margin, and the 206% growth in referral traffic tells you everything you need to know about its trajectory as a marketing channel.

The businesses building their AI search strategy now — not waiting for the dust to settle — will hold the advantage as this channel continues to grow.

Next Steps

This week: Set up AI search tracking in your Google Analytics. Create a custom channel group that separates ChatGPT, Gemini, Claude, Perplexity, and Copilot referral traffic. Look at both volume and conversion rate by platform. This data will tell you exactly where to focus your optimisation effort. If you need help setting this up, our guide on tracking your brand’s visibility in LLMs walks through the process step by step.

Next 30 days: Audit your website content through an AI lens. Are your product or service pages clear enough for an AI platform to understand exactly what you sell, who it’s for, and why you’re the best option? Run an AI Search Optimisation audit to identify your biggest gaps — both on your own site and across the third-party sources that ChatGPT draws from. Review your positioning and customer personas with fresh eyes. The businesses with the clearest positioning will win in AI search.

Next 90 days: Build a comprehensive AI Search Optimisation strategy that covers your own content, your third-party presence (media coverage, reviews, user-generated content), and your digital PR priorities. Think beyond keywords — map the topical areas you want to own and the types of prompts your ideal customers are likely to use. If you’d like expert guidance, request a free marketing review from Exposure Ninja and we’ll show you exactly where your brand stands in AI search results — and what to do about it.

In Conclusion

The Semrush data makes one thing undeniable: ChatGPT is no longer a novelty. It’s a search channel — one that’s growing fast, influencing purchase decisions, and sending meaningful referral traffic to websites across every industry.

The fact that 21% of that traffic goes to Google isn’t a weakness in the ChatGPT story. It’s proof that these platforms work together, that the buyer journey is now multi-platform, and that your SEO foundation matters more — not less — in a world where AI is reshaping how people discover and evaluate brands.

The businesses that will benefit most from this shift are the ones taking action now: tracking their AI referral data, auditing their visibility across platforms, building content strategies that go beyond keywords, and investing in the third-party presence that AI platforms rely on when making recommendations.

As Charlie puts it: “We need to optimise for all of the channels where we know that our customers are having touch points.”

ChatGPT is one of those channels. The data says so. The question is whether your strategy reflects it.

Is your marketing underperforming?

Request a website and marketing review and our team will tell you how to improve your marketing.

Request a Free Review