AI chatbots have opened up a whole new customer acquisition funnel. And marketers now have to grapple with the exact same questions of resource allocation that dictate search engine budgets: prioritizing optimization for organic discovery, or paid advertising.
Exploding Topics wanted to find out how consumers are really researching products online, to help marketers and brand strategists make more informed decisions about their AI optimization strategy. So in our biggest survey to date, we asked over 2,300 people in the U.S. how they use AI in their purchasing journeys.
The picture that emerged was one of an AI gatekeeper quietly swinging consumer purchase decisions.
Chatbots have become a big part of the average consumer’s product research, and customers are discovering new brands through AI. Consumers are researching local and large-scale businesses alike, and their AI usage is increasing over time. Almost two-thirds of the AI users we surveyed have at least partially replaced product-related Google searches with AI conversations.
But as well as being an avenue for brand discovery, AI chatbots are also talking would-be customers out of purchases. More than half of AI users have decided not to purchase something on the basis of information from a chatbot.
At the same time, more than 4 in 10 people actively dislike AI chatbot ads, and the average consumer’s opinion of a brand is slightly more likely to worsen than improve after seeing an ad for it. So chatbot ad space is not necessarily the easy fix to AI visibility issues.
Instead, our results show the real and present importance of Generative Engine Optimization (GEO) efforts. Keeping on track of your AI brand presence isn’t just a luxury for securing extra customers, it’s an increasingly essential prerequisite for ensuring your sales aren’t getting intercepted.
Fast facts
Among all consumers:
- Roughly 1 in 5 would pick an AI chatbot if they could only use one product research tool
- Almost half (47.54%) sometimes consult an AI chatbot to get seller information before making a purchase
- 74.15% would be at least somewhat less likely to buy a product if an AI chatbot flagged mixed or negative reviews
Among all AI users:
- 65.01% have at least partially replaced Google searches with AI chatbots for product research
- 71.63% feel neutral or negative toward ads in AI chatbots
Among those who use AI at least weekly:
- 73.58% have bought a product based on an organic AI recommendation
- 81.24% believe AI chatbots are capable of recommending the best brands and products
- Almost 7 in 10 (69.7%) have decided not to buy something on the basis of AI advice
AI chatbots are the new influencers
Organic AI mentions have an increasingly decisive impact on consumer activity.
41% of consumers have made a purchase based explicitly on an AI chatbot recommendation. Among all consumers who use AI at least occasionally, that figure rises to 57.51%.
And the picture is even more stark among AI power users. 80.45% of consumers who use AI multiple times per week have bought an AI-recommended product at least once, with more than half (54.95%) doing so multiple times.
Crucially, the persuasive power of AI chatbots works in both directions. 57.5% of AI users have decided not to buy something based on information provided by an AI chatbot.
When you filter further to only include the consumers who named AI chatbots as one of the sources they use when making a purchase online, 80.93% of people have decided against a purchase on the basis of AI guidance. 42.81% have done so multiple times.
Looking at all consumers again, 74.15% would be at least somewhat less likely to buy a product if an AI chatbot flagged mixed or negative reviews. Among current AI users, 85.31% would rethink a purchase if a chatbot revealed unfavorable reviews.
Knowing that AI chatbots can dissuade users from purchases almost as much as they can persuade them underlines the reality that GEO is no longer optional. Marketers who aren’t optimizing for AI aren’t just missing out on potential acquisitions, they may be having would-be purchases actively intercepted by chatbots.
AI mentions have tangible brand impacts
AI recommendations can do more than influence a single purchase. They can introduce brands and products to whole new audiences.
59.27% of all AI users have discovered a brand or product they were previously unaware of through an AI chatbot recommendation. More than a quarter (25.79%) have “frequently” discovered new brands and products.
(These findings tally with Semrush’s buyer journey study from March, which found 43% of consumers have discovered a new brand through an AI tool. Across all consumers, our survey puts the same figure at 43.14%.)
Among consumers who are actively using AI for product research, only 9.38% have never discovered a new product or brand from a chatbot recommendation.
Brands stand to benefit massively when AI chatbots introduce them to new customers. More than half of all consumers believe AI chatbots are capable of recommending the best products and brands.
And when you focus on the consumers who currently use AI, 70.62% believe that chatbots can recommend the best brands and products. That rises above 80% for those using AI at least weekly.
In other words, a significant majority of the vast and growing AI chatbot user base believes that the technology can provide valuable and actionable brand/product recommendations. Looked at through this lens, optimizing for AI brand mentions seems increasingly non-negotiable for marketers.
As well as discovering new brands through organic mentions, consumers are also using AI chatbots to actively research brands they are already aware of.
Almost half (47.54%) of all consumers now at least occasionally consult AI chatbots to get information about a company or seller before making a purchase decision. Brands now have to account for the reality that there is a near 50-50 chance that a would-be customer might consider running an AI “background check.”
Excluding AI non-users, 65.42% of consumers report sometimes consulting an AI for company information before committing to a purchase. Among these AI users, 32.68% “frequently” consult AI chatbots for information about a seller.
Sure enough, most of the consumers who have ever researched a company using an AI chatbot have done so recently. 48.97% have researched a company using AI within the last week, and a further 20.3% have done so in the past 2 weeks.
So what will AI say about your brand when your customers ask? Coupled with our findings on the influential role of chatbots as potential purchase interceptors, this looks to be an increasingly fundamental question for marketers to answer.
The power of AI chatbots for local businesses
Local, physical businesses might think that AI is of limited relevance to them. As it turns out, customers think very differently.
In fact, 87.41% of people who use AI for company/seller research would at least occasionally consult an AI before hiring a local business.
Almost a third (31.48%) of consumers who use AI for company research would always consult AI chatbots before hiring a local business.
In many cases, consumers believe AI chatbots are better at giving advice on local businesses.
42.01% of people who use AI to research companies/sellers think that AI gives advice on local businesses better than it recommends large-scale companies. A further 43.7% believe AI handles local business advice just as well as it handles big brands.
For business owners, the lesson is clear: regardless of your scale and scope, many consumers will use AI to research your company. Taking control of the responses they receive can have a tangible impact on your bottom line.
Consumers turn to AI en masse for product research
Discussion about consumer adoption of AI is no longer a speculative topic: chatbots have already been adopted at scale across most demographics.
33.23% of people use AI chatbots like ChatGPT, Gemini, Claude, and Perplexity multiple times per week. A further 15.41% use chatbots weekly.
There is still a significant minority (29.96%) who do not use AI chatbots at all. But AI non-usage is primarily confined to the over-60s, among whom 60.05% never use AI chatbots. Only 13.61% of all consumers aged 18-60 report never using AI chatbots.
Likewise, 65.5% of consumers aged 18-60 use AI chatbots at least once per week.
And when a consumer is a regular AI user, it’s likely that they use AI for product research.
55.13% of respondents who use AI at least once per week also said they use AI chatbots as a source of information when making a purchase online. Among these regular users, AI is a more popular product research tool than friends and family (49.28%), social media (48.38%), retail websites (44.42%), and YouTube (43.35%), though search engines and reviews still lead.
Across all demographics, including AI non-users, a more modest 29.09% of consumers report using AI chatbots for product research. But more than 60% of respondents expect they will be using AI for product research more frequently in 2 years’ time if the technology continues to advance at a similar pace. So marketers should be treating GEO as a permanent and increasingly important strategic pillar.
AI is also steadily growing in prominence as a first port of call for product research.
Excluding complete AI non-users, 12.72% of consumers now start their online purchasing journey with AI chatbots. Search engines (29.89%) and online reviews (28.13%) are still more common starting points, but AI is the third-most-common place for consumers to begin their research.
AI users with upper-middle household incomes are more likely on average to start their purchasing journeys with an AI chatbot. 25.12% of AI users with a household income of $125,000-$199,999 turn to AI chatbots first, narrowly beating online reviews (24.13%) and only trailing search engines (35.58%).
Interestingly, the percentage of AI users who start online purchases via chatbots drops back down to 10.19% among the very highest earners ($200,000+ household income). That’s comparable to lower earners; only 8.8% of users with household incomes from $0-$124,999 begin purchasing journeys by consulting chatbots.
Nonetheless, consumers aren’t just starting their product research with AI. Almost 1 in 5 people (19.43%) would choose an AI chatbot if they could only pick 1 source before making a purchase.
Filtering out those who don’t use AI chatbots at all, more than a quarter of consumers (27.37%) would opt for chatbots as their sole source of product research if made to choose. Meanwhile, 44.39% of AI “power users” (multiple uses per week) would select AI chatbots as their only product research tool, versus 26.29% for search engines and 22.59% for reviews.
AI chatbots challenge Google supremacy
Consumers are adopting AI chatbots for product research with such enthusiasm that tools like ChatGPT are gradually beginning to eat into Google’s near-monopoly.
More than 1 in 5 AI users have completely replaced their product-related Google searches with AI chatbots. And 65.01% of consumers who use AI have at least partially swapped out Google for chatbots.
AI users aged 45-60 are most likely to have phased out Google entirely for product research, with 39.04% reporting that AI chatbots have “completely” replaced the search engine. By contrast, 63.36% of AI users aged over 60 report that they have not replaced any product-related Google searches with AI chatbot conversations.
Despite many users swapping some Google searches for AI queries, the average consumer has not reduced their overall usage of Google for purchase research in the past year. Almost half (47.6%) of AI users have roughly maintained their level of Google usage in the last 12 months, while 23.01% report using the search engine much more.
So AI chatbots are very much complementing Google searches rather than replacing them, even though lots of consumers have migrated specific queries away from the search engine in favor of AI. The best marketing strategies consequently need to encompass both SEO and GEO.
Shoppers favor AI responses over AI ads
Until recently, most of the main AI chatbots have been ad-free experiences. But while companies like OpenAI have been content to run at large losses for the sake of user acquisition, attention has begun to turn to more effective monetization beyond just premium and enterprise plans.
In the US, ads have been active in ChatGPT since February, with the beta self-serve advertising platform launching with no minimum spend in early May. 42.22% of consumers have now seen an ad in an AI chatbot; 35.85% have not seen an ad, while a surprisingly high 21.94% aren’t sure if they have seen an ad or not.
Consumer attitudes here are moving fast. When Exploding Topics surveyed consumers in April, 48.35% expected shopping features and ads to improve AI responses. That survey ran while chatbot ads were still in limited testing; this one ran after months of wider rollout. Watching sentiment shift from optimism in theory to skepticism in practice is exactly why brands need to track this space continuously.
The level of uncertainty among consumers as to whether or not they have seen a chatbot ad speaks to an interesting point. When a consumer is using an AI chatbot for product research, the line between an organic recommendation and a paid advertisement can quickly become blurred from a user’s standpoint.
Perceived ambiguity between organic and paid product suggestions helps to explain why among the 41.63% of consumers who dislike chatbot ads, 66.05% say that ads make them doubt the integrity of AI responses. Integrity concerns are by far the biggest reason for consumers to dislike chatbot ads:
Only around a quarter (26.81%) of people who dislike chatbot ads specifically say that ads are insufficiently separated from organic AI content. But even when adverts are clearly marked, their mere inclusion is evidently troubling a significant number of consumers when it comes to trust.
Concern about the integrity of AI chatbot ads climbs with age. 68.86% of surveyed users aged 60+ say ads make them doubt the integrity of the AI responses, versus 59.13% of surveyed 18-29-year-olds. However, integrity concerns are comfortably the most common complaint about chatbot ads across all age groups.
An issue from a marketer’s point of view is that ads in AI chatbots have the potential to make consumers doubt the entire response, organic or otherwise. But brands can only work on things within their control, and it seems abundantly clear that potential customers value organic AI mentions more than paid chatbot ads.
The impact of AI chatbot ads
Not all consumers dislike ads in AI chatbots. 38.09% of people are neutral, and 20.28% like the ads.
We asked the neutral/like group what they like about ads, and presented them with a list of options. More than half (54.85%) said that chatbot adverts acted as an extra source of information.
But from a marketing standpoint, the more pertinent question is whether AI chatbot ads are effective. And our data shows that running AI chatbot ads is potentially more likely to harm your brand than help it.
20.36% of consumers would have a better opinion of a product after seeing an ad for it in an AI chatbot platform. But for 23.12% of consumers, the ad would trigger a worse opinion of the product.
It’s important to note that the majority (56.52%) of consumers said that ads would have a neutral impact on their opinion of a product, or that it would depend on the content of the ad. So a well-constructed chatbot ad still has the potential to have a positive impact on consumers. But marketers are effectively starting with a handicap in their efforts to achieve strong ROI.
In fairness, 54.04% of people who have seen ads in chatbots self-report clicking at least once. But that leaves over a third (35.85%) who report never clicking, while a further 10.11% are unsure if they have clicked.
Overall, ads in ChatGPT are also less likely to lead to a click than ads in Google SERPs, per self-reported consumer preferences. 24.88% of consumers say they would be more likely to click on a ChatGPT ad than a Google ad, but 45.42% say they would be more likely to click on a Google ad.
However, when you only consider consumers who have seen an ad in an AI chatbot, the picture flips. 48.94% of consumers who have been served AI ads before say they are more likely to click on a ChatGPT ad, while only 30.64% say they are more likely to click on a Google ad.
So AI chatbot ads are certainly not useless, and clickthrough may end up comparable with or even better than Google ads as consumers begin to see them with more frequency. But with AI, much like with Google SERPs, optimizing to appear in organic results is likely to produce the best sustainable and long-term ROI.
And as a final point, marketers need to consider whether their potential customers are likely to start seeing more AI chatbot ads any time soon. ChatGPT, for example, only serves ads on its Free and Go plans. Plus, Pro, and business accounts remain ad-free. So for many B2B products, there’s a risk that the main target market won’t see the ads at all.
Put GEO at the heart of your AI chatbot strategy
The results we gathered show how consumers are using AI chatbots in their buying journeys. They are using and trusting AI for brand discovery, product recommendations, and pre-purchase checks, and advice from a chatbot can block a purchase as easily as it can lead to one.
The wider rollout of ads in leading chatbots like ChatGPT may seem like it presents a simple avenue into the bottom-of-the-funnel conversations that high-intent users are having with AI agents. But advertisers are fighting against a strong base level of chatbot ad skepticism, and can even end up doing more harm than good to a brand’s reputation.
The most effective and reliable route to harness the ever-growing impact of AI on product research is a comprehensive Generative Engine Optimization strategy. GEO is how you make sure consumers are finding your brand organically, exactly when you want them to be.
In practice, that means three things: know what AI chatbots currently say when consumers ask about your brand and category, fix the gaps in the sources those answers draw from, and track your share of AI recommendations the same way you track rankings.
Semrush provides the best suite of tools for monitoring and improving your brand’s LLM presence, at both the enterprise and SMB scale.
Semrush Enterprise AIO lets you monitor more than 289 million relevant LLM prompts, tracking your organic mentions and finding gaps in your visibility profile. You’ll have instant access to data-driven recommendations on how to show up in more AI queries, encompassing technical backend changes and content fine-tuning.
Your bespoke package will be tailored to your industry and goals, and you’ll have access to comprehensive analytics for benchmarking against competitors. You’ll also be able to share GEO data across all of your teams via API.
For smaller businesses, individual marketers or freelance SEO specialists and brand strategists, the Semrush AI Visibility Toolkit is the best GEO tool on the market. With one-click AI visibility analysis reports, prompt gap analysis, and continuous share-of-voice tracking, you can bring all the rigor of elite SEO into the new era of AI consumer product research.
More than 80% of AI power users are already buying products (or not buying them) based on AI recommendations. Start winning these customers, and book a Semrush Enterprise AIO demo today.
Methodology
We surveyed 2,338 adults in the US in July 2026. The margin of error is plus or minus 2 percentage points at a 95% confidence level for full-sample statistics; filtered segments carry wider margins.
Respondents were drawn from a mix of regions and economic backgrounds. The median household income fell in the $75,000-$99,999 band.
The survey divided respondents into 4 age bands: 18-29, 30-44, 45-60, and 60+. There were at least 300 respondents in each band.
We filtered our data in various ways in order to get the most comprehensive picture. As such, the denominator sometimes differs between statistics cited in this report. This is a non-exhaustive list of the segments we examined, and how they are referenced in the report:
- “Consumers”: All survey respondents
- “AI users”: Respondents who reported using AI at any frequency
- “Regular AI users”: Respondents who reported at least weekly AI usage
- “AI power users”: Respondents who reported using AI multiple times per week
