For more than two decades, one of the most important moments in customer acquisition happened when someone opened Google and typed a search. Businesses built websites, advertising programs and entire marketing strategies around being visible at that moment. Search still plays an enormous role in how customers find companies, but new consumer research shows that it is no longer the only place where those decisions begin.
Customers are increasingly taking questions they once would have broken into a series of Google searches and bringing the entire problem to ChatGPT and other AI-powered tools. Instead of searching, opening several websites, returning to the results and searching again, they can describe what they need, ask follow-up questions, compare alternatives and narrow their choices inside one conversation. The change isn’t simply from one search engine to another. It is a change in how people research a decision in the first place.
One in Four Consumers Already Uses AI as a Top Research Source
Adobe’s 2026 consumer research provides one of the clearest indications of how quickly this behavior is developing. Traditional search remains the most commonly used source for researching information, making purchase decisions and finding recommendations, with 48.4% of consumers identifying search engines as one of the sources they use most often. But 24.9% now identify AI-powered platforms such as ChatGPT, Gemini and Copilot as a top source for that same research. AI-powered platforms have already moved ahead of company or brand websites at 21% and online review sites at 16%.
That is significant because ChatGPT isn’t simply being used to answer trivia or summarize information. Consumers are bringing decisions into these conversations. Adobe found that among people already using AI-powered platforms for research, 42% always or frequently rely on AI assistants, conversational platforms or AI-powered search as a primary source for advice, shopping or troubleshooting. Nearly half of consumers also said they would use AI for personalized product recommendations.
The numbers become even more pronounced when looking specifically at people who already use ChatGPT. In separate Adobe research examining how people use ChatGPT for search, 77% of U.S. ChatGPT users said they use it as a search engine, 24% said they go to ChatGPT first, and 36% said they had discovered a new product or brand through ChatGPT. That puts ChatGPT directly into a part of the customer journey that businesses have historically depended on search engines to deliver.
A Conversation Can Replace a Series of Searches
The difference between searching and asking ChatGPT becomes clearer when considering how people actually research complicated decisions. Traditional search often requires the customer to translate a problem into multiple individual queries. Someone may begin broadly, learn something from the results, refine the search, visit several websites and continue repeating the process until enough information has been assembled to make a decision.
A ChatGPT conversation can keep that context from one question to the next. The customer can explain the original problem, add requirements, challenge a recommendation, ask for alternatives and narrow the decision without starting over each time. Instead of the customer assembling information from a collection of search results, the conversation itself becomes part of the research process.
That distinction matters to businesses because it can change when companies enter the customer’s consideration set. A business doesn’t necessarily have to wait until someone knows enough to search for its product category, service or brand by name. It can become relevant while the customer is still trying to understand the problem and determine which options deserve consideration.
Being Recommended in ChatGPT Can Influence What Happens Next
One of the challenges in measuring this shift is that ChatGPT doesn’t always receive credit for the website traffic or purchase that follows. A customer can research a decision in ChatGPT, learn about a company and later search Google for that company’s name rather than clicking a link directly from the conversation. Traditional analytics may record Google as the source even though the company’s inclusion in the ChatGPT conversation helped create the visit.
Similarweb’s research into the downstream impact of AI visibility provides evidence that this is already happening. Using U.S. desktop clickstream data, Similarweb followed users after they received brand recommendations in ChatGPT. It found that brands recommended by ChatGPT were 2.5 times more likely to receive a website visit within seven days than competitors that weren’t recommended in the same category. Even more importantly, 55.9% of that downstream traffic arrived through branded search rather than through a visible AI referral.
That finding changes how businesses should think about AI visibility. Looking only at the number of visitors arriving directly from ChatGPT can substantially understate its influence. If someone asks ChatGPT for help, encounters a company during the conversation and later searches Google specifically for that company, Google may get the attribution while ChatGPT played an important role in determining which company the customer searched for.
ChatGPT Traffic Is Still Small, but the Customers Arriving From It Matter
It is important not to overstate how far this shift has progressed. Direct traffic from ChatGPT remains much smaller than traffic from traditional search. A peer-reviewed study published in Marketing Science analyzed 12 months of first-party data from 973 ecommerce websites representing more than $20 billion in revenue and more than 50,000 transactions originating from ChatGPT referrals. The researchers found that organic ChatGPT referrals accounted for less than 0.2% of overall traffic in the dataset, making the channel roughly 200 times smaller than Google organic search.
But the same study found evidence that makes the relatively small traffic volume difficult to dismiss. ChatGPT referral traffic produced higher conversion rates and revenue per session than paid social, although it remained below the other traditional channels examined in the study. Performance was also stronger for more complex product categories, where customers have more information to evaluate before making a decision. The researchers found conversion rates from ChatGPT referrals increasing over time, with their cross-website analysis supporting growing consumer proficiency with LLMs as a possible explanation.
That combination is more important than a simple comparison of traffic volume. Search has had decades to become embedded in consumer behavior and business marketing. AI-assisted research is much newer, yet consumers arriving from ChatGPT are already demonstrating measurable commercial behavior. Businesses therefore shouldn’t evaluate the opportunity solely by asking how much traffic ChatGPT sends today. The more important question is how much influence AI conversations are beginning to have over what customers consider and eventually buy.
Even Google Search Is Becoming More Like an AI Conversation
The movement toward AI-assisted research isn’t occurring only outside Google. Google itself is increasingly incorporating AI-generated answers into search. Similarweb’s 2026 Generative AI Landscape report found that more than 40% of U.S. Google searches now trigger an AI Overview. Similarweb also reported that Google search queries have become 5.4% longer since the introduction of AI Mode, another indication that search behavior itself is becoming more conversational.
This makes the change larger than a competition between Google and ChatGPT. Customers are becoming accustomed to asking more complete questions and receiving synthesized answers rather than simply choosing among a list of links. Whether that interaction begins inside ChatGPT, Google or another AI-powered experience, businesses increasingly need to consider whether they are represented inside the answer, not only where their webpage ranks underneath it.
The Customer Journey Is Changing Before the Website Visit
For years, businesses could focus much of their digital customer acquisition strategy on what happened after someone reached Google: which keyword they searched, where the company ranked, which advertisement appeared and which landing page received the click. AI conversations add another layer before that website visit. A customer can now conduct substantial research, encounter companies, eliminate alternatives and form preferences before ever reaching a company’s website.
That doesn’t make traditional search irrelevant. In fact, the Similarweb data shows that search can remain an important next step after an AI recommendation. What changes is the role search plays in the journey. The customer arriving through a branded Google search may no longer be discovering the company there. The discovery may have happened several minutes, hours or days earlier inside an AI conversation.
For businesses, that makes visibility inside those conversations increasingly important. If customers are using ChatGPT to determine which products, services and companies deserve consideration, businesses need to understand whether they are part of those answers, which competitors are appearing instead and what information is influencing the recommendations. With paid advertising now becoming part of ChatGPT as well, companies have both organic and paid ways to compete for visibility during these decisions.
The shift from search to AI isn’t a future scenario waiting for consumers to change their behavior. The behavior has already started. Search remains massive, but customers are adding a new step to the way they research and make decisions, and in some cases that step is happening before Google ever enters the journey. For businesses that have spent years competing to be found when customers search, the next challenge is making sure they can also be found when customers ask.