AI search statistics 2026: what the latest data says about spending and trust

AI search has finished being a debate. It is a measurable consumer channel that is influencing what people buy, who they trust to recommend it, and where the money lands. This is our running collection of AI search statistics, covering Google AI Overviews, ChatGPT and other generative AI search tools. The numbers below are not from the agency-pitch-deck side of the conversation. They come from named research firms, public earnings, official platform data and reputable third-party studies, all published in 2025 or 2026.

This page is current as of August 2026, and we update it every month as new AI search data is published. The pattern across categories is the same. Adoption is faster than most marketing teams expected. Conversion behaviour is meaningfully different from traditional search. And brand discovery is moving away from the places most businesses are still optimising for. If you are reading this in the second half of 2026, the window to be early is closing fast in some categories and already shut in others.

New in the August update: Google's own adoption figures from I/O, Pew's read on how many Americans use chatbots and what they use them for, Semrush's study of 126 million AI search prompts and what it says about which brands the platforms can see, and the first hard evidence that agent-led checkout is running at scale in some markets. There is also a new section on why visibility on one platform tells you almost nothing about visibility on the next.

Here is what the latest data says.

The key AI search statistics for 2026 at a glance

If you only take a handful of numbers from this page, take these. Each one is sourced in full further down.

• Google AI Overviews now reach 2.5 billion monthly users, and AI Mode passed 1 billion in its first year.
• 49% of US adults use AI chatbots, and searching for information is the most common thing they do with them.
• 44% of AI-powered search users say it is now their primary, preferred source for buying decisions.
• AI search traffic converts at 14.2%, against 2.8% for Google organic.
• 73% of B2B buyers now use AI tools in their purchase research.
• When an AI Overview appears, users click a source inside it just 1% of the time.
• Of more than 1,200 brands tracked across 4 AI platforms, only 36 held top-100 visibility on all of them every month.
• 45% of marketing leaders cannot measure how their brand shows up in AI answers, and only 9% have the tools to track it properly.

How has AI search changed consumer buying behaviour in 2026?

44% of AI-powered search users say it is now their primary and preferred source for buying decisions.

McKinsey's "New front door to the internet" research found that 44% of users put AI search ahead of traditional search engines (31%), retailer or brand websites (9%) and review sites (6%). Around 50% of consumers polled now intentionally seek out AI-powered search engines as their top digital source for buying decisions.

Source: McKinsey, New front door to the internet, November 2025

$750 billion in US revenue will funnel through AI-powered search by 2028.

McKinsey projects three quarters of a trillion dollars of US consumer spending will move through AI search surfaces within 3 years. The same report identifies 40 to 55% of consumers in the top spending sectors (consumer electronics, grocery, travel, wellness, apparel, beauty and financial services) as already using AI-based search to make purchasing decisions.

Source: McKinsey, New front door to the internet, November 2025

Google AI Overviews reach 2.5 billion monthly users. AI Mode passed 1 billion in its first year.

Sundar Pichai put both numbers on stage at Google I/O in May 2026. AI Overviews had crossed 2 billion the previous July, so the surface added another half a billion people in under a year. AI Mode, the conversational version of Search, went from launch to a billion monthly users in twelve months, with queries more than doubling every quarter. Google repeated both figures in June, when it started testing a Search Console control that lets site owners keep their pages out of generative Search features.

Source: Google, Search at I/O 2026, May 2026

49% of US adults use AI chatbots, and searching for information is what they use them for most.

Pew Research surveyed 5,119 US adults in February and published the findings in June. Chatbot use has gone from 33% in 2024 to 49%, with about a quarter of adults using one every day. Asked what they do with them, 42% said searching for information, which puts search ahead of work tasks, entertainment and image editing. Separately, 60% of US adults say they read the AI summaries at the top of search results. Half the country is now taking answers from a system that decides on its own which brands to name.

Source: Pew Research Center, Americans and AI 2026, June 2026

AI search traffic converts at 14.2%. Google organic converts at 2.8%.

AI referral traffic is roughly 5 times more valuable per session than traditional organic search. Claude visitors convert at 16.8%, ChatGPT at 14.2%, Perplexity at 12.4%. The same study found AI visitors spend 68% more time on websites than traditional search visitors.

Source: Superprompt, AI Search Traffic Conversion Rates, 2025

42% of consumers who use AI rely on it always or frequently as their primary source for advice, shopping and troubleshooting.

Adobe's 2026 AI and Digital Trends report (a global survey of 4,000 consumers conducted October to November 2025) found that among customers already using AI-powered platforms, reliance is high. Around a quarter of consumers overall now cite AI platforms like ChatGPT as their top research tool, ahead of brand websites and online reviews.

Source: Adobe, 2026 AI and Digital Trends Consumer Report

How is AI search changing retail and DTC spending?

Retail is the category where AI has already shown up in the revenue line, not the projection. The 2025 holiday season was the first year where AI traffic outperformed traditional traffic on conversion, and the numbers are not subtle.

AI referral traffic to US retail sites surged 693% during the 2025 holiday season.

Adobe Analytics tracked a 693.4% year-on-year increase in traffic from generative AI sources to US retail websites between 1 November and 31 December 2025. The earlier benchmark, the same metric in March 2025, sat at a 1,200% YoY jump. AI is now an established source of retail traffic, not an emerging one.

Source: Adobe, 2025 Holiday Shopping Report, January 2026

AI traffic to US retail sites is up 1,324% since October 2024. Travel is up 2,215%.

Adobe's numbers, published alongside Semrush's June study, run to May 2026 and show travel outpacing every other category by a wide margin. Retail is the category people talk about, but the operators with the most to gain right now are lodges, tour companies and anyone else selling a trip.

Source: Adobe data via Semrush AI Visibility Index 2026, June 2026

AI-driven shoppers converted 54% better than non-AI traffic on Thanksgiving and 38% better on Black Friday.

Adobe found that AI referrals moved from lagging to leading on conversion during the 2025 holiday peak. Revenue per visit from AI sources was up 254% year-to-date. Shoppers arriving from AI assistants were 33% less likely to bounce immediately than other traffic.

Source: Adobe, 2025 Holiday Shopping Report

Amazon's shopping assistant reached 250 million monthly active users and drove $12 billion in incremental annualised sales in 2025.

Amazon's in-platform AI shopping assistant, then branded Rufus, grew its user base 149% year-on-year and total interactions 210%. In November 2025, Amazon rolled out agentic features that let it autonomously buy products on behalf of customers, including auto-buy at target prices. In May 2026, Amazon retired the Rufus name and folded the assistant into Alexa for Shopping, a single agent built into the main Amazon search bar.

Source: Fortune and PYMNTS, November 2025

24% of consumers are comfortable with AI agents shopping for them. Among Gen Z, that number rises to 32%.

Salesforce's 2025 research on AI agent adoption shows roughly a quarter of consumers are already comfortable handing routine purchases to an AI agent. The rate is meaningfully higher in younger demographics. Google's Universal Checkout Protocol launched in February 2026 with Etsy and Wayfair, letting users buy without leaving Gemini or AI Mode.

Source: Salesforce, 2025

A third of consumers expect at least 10% of their purchases to be AI-driven within a year. A quarter say they will never hand over the card.

Checkout.com's Agentic Commerce 2026 research found demand and trust moving at different speeds. 33% of consumers expect AI to be driving at least a tenth of their purchases inside twelve months, while 24% say they will never delegate a purchase to an agent and 27% trust no organisation to run one on their behalf. On the merchant side, 72% agree that shoppers will adopt agent-led buying faster than most businesses are ready for. The distance between those two numbers is the opening for brands who get their product data, pricing and returns terms legible to machines early.

Source: Checkout.com, Agentic Commerce 2026, June 2026

Alipay processed more than 120 million AI-agent payments in a single week.

NielsenIQ's mid-year read on global commerce, published at the end of July, puts agent-led checkout in the category of working infrastructure rather than pilot project. Ant Group reported over 120 million AI Pay transactions in the week of 5 to 11 February 2026. Morgan Stanley's estimate for the US puts agent-driven spending at $190 to $385 billion, roughly 10 to 20% of e-commerce, by 2030. On this particular shift, the East is a long way ahead of where most Western retail teams think the market is.

Source: NielsenIQ, The Commerce Revolution, July 2026

AI agents could mediate $3 to $5 trillion of global consumer commerce by 2030.

McKinsey's broader projection on agentic commerce puts the addressable share of consumer spending influenced or executed by AI agents within 5 years at multiple trillions of dollars. That is the upper bound of what is at stake. The lower bound is just whether your product description, schema and reviews are clear enough for an AI to pick you up.

Source: McKinsey, Superagency in the workplace

How are B2B and SaaS buyers using AI search to choose vendors?

B2B buying has shifted to AI faster than retail, partly because the buyer is already accustomed to multi-source research and because the average B2B deal justifies the time investment of asking ChatGPT a hard question. The discovery layer of the B2B funnel, the bit where buyers build a shortlist before talking to a vendor, has already moved.

73% of B2B buyers now use AI tools in their purchase research.

A March 2026 analysis of 680 million AI citations by Averi found that almost 3 quarters of B2B buyers use ChatGPT, Perplexity, Gemini or similar tools as part of their vendor research. A separate Forrester finding put the number at 94% for buyers using AI to build their shortlist before they ever contact a vendor.

Source: Averi research via PR Newswire, March 2026

66% of B2B buyers now use AI specifically for supplier research.

TraxTech's 2025 supply chain study found that two-thirds of B2B buyers are using AI to research suppliers, evaluate vendors and compare options before any human conversation happens. The implication is that your supplier discoverability is no longer a function of trade shows, directories or paid search. It is a function of how AI describes you.

Source: TraxTech, AI in Supply Chain, 2025

Twice as many B2B buyers now name generative AI as their most meaningful research source than any other channel.

Generative and conversational AI now outranks vendor websites, product experts and sales representatives as the most useful research source named by B2B buyers. Vendor-controlled surfaces are no longer the first stop. The summary of you that lives inside an AI model is.

Source: Forrester and Averi, via TraxTech

AI agents will command $15 trillion in B2B purchases by 2028.

Gartner projects that B2B purchasing mediated by AI agents will reach $15 trillion within 3 years, a figure that includes everything from automated procurement to agent-led RFP responses. By the end of 2026, Gartner expects 40% of enterprise apps to include conversational AI agents.

Source: Gartner via Digital Commerce 360, November 2025

Roughly 30% of consumer ChatGPT usage is already work-related, and that is before you count Enterprise plans.

OpenAI's February 2026 publication of usage data, based on 1.8 million conversations analysed in partnership with NBER and Harvard economist David Deming, confirms that ChatGPT is functioning as a B2B research channel by stealth. The 30% figure excludes ChatGPT Enterprise and Team usage entirely. Real B2B exposure is higher.

Source: OpenAI Signals, February 2026

Only 22% of marketers are currently tracking AI visibility and traffic.

Digitaloft's 2025 survey of marketers found a striking gap. Even with B2B buyers already using AI to shortlist vendors, fewer than a quarter of the teams who would be affected are measuring it. Only 25.7% plan to develop content specifically for AI citations. This is the gap most B2B brands have not yet noticed.

Source: Digitaloft AI in SEO Statistics, 2025

How is AI search changing how people plan and book travel?

Travel is the category where the buyer behaviour shift is sharpest. Trip planning is exactly the kind of multi-step, comparison-heavy, opinion-driven research task that AI assistants do well. Consumers have noticed. Booking platforms have noticed. Most hotels and lodges have not.

Around 40% of travellers globally are already using AI tools for trip planning and booking.

Statista's 2025 industry overview puts global traveller adoption of AI-based planning tools at roughly 4 in 10. Around 80% say they are open to using AI for planning, booking and experiencing their holiday, even if they have not done so yet. Generating trip itineraries is the single most common AI use case among travellers.

Source: Statista, AI in travel and tourism, 2025

Search engines fell from 51% to 36% of travel research between late 2024 and the second half of 2025. Generative AI more than doubled in the same window.

Generative AI platforms grew from 6% to 15% of traveller research activity in less than a year. Search engines lost 15 points of share. For a category where the discovery journey often runs months before the booking, this is the kind of shift that compounds quickly. Travellers are forming opinions about destinations and operators inside AI tools long before they get near a booking engine.

Source: Phocuswire, AI developments in travel, 2025

ChatGPT referrals to travel websites grew 300% in just a few months.

Semrush data showed referral traffic from ChatGPT to travel sites tripling in the first part of 2025. The travel category is among the fastest-moving in terms of AI referral growth, partly because the typical research session lasts long enough for a user to click through to the source.

Source: Semrush via Hospitality Net, 2025

Up to 60% of Gen Z and Millennial travellers use AI during trip exploration. Almost 1 in 4 American Gen Z and Millennial travellers specifically use ChatGPT for planning.

Younger travellers are not just using AI to research destinations. They are using AI to surface options they would not have found through traditional search at all. For lodges, boutique operators and DMCs whose discoverability has historically depended on third-party aggregators, this changes the front of the funnel entirely.

Source: Statista via Mindful Ecotourism, 2025

83% of travellers say they are more likely to book when AI-enhanced services are offered.

Travel buyers are not just open to AI in the planning phase. They reward operators that embed AI into the buying experience itself. Combined with the 65% of travel industry leaders who name chatbots and virtual assistants as the most significant generative AI application in their category, the demand and supply sides are aligned. The question is which operators move first.

Source: Statista via Mindful Ecotourism, 2025

How is AI search changing how people manage money and choose financial products?

Finance is the quietest of the 4 categories on the surface, but the consumer-side adoption is already at scale. Personal money management, investment research and financial planning are now mainstream AI use cases, particularly in the UK. The B2B and product side is moving alongside it.

56% of UK adults (around 28.8 million people) have used AI in the past 12 months to help manage their money.

Lloyds Banking Group's 2025 research is the largest UK consumer dataset on AI and personal finance to date. Among the people using AI to manage money, ChatGPT is the dominant tool: 6 in 10 of them are using it specifically. AI for money management is no longer a fringe behaviour confined to tech-forward users.

Source: Lloyds Banking Group, August 2025

37% of AI users are using it for investment research and recommendations. 39% are using it for future financial planning, including pensions.

Lloyds found that financial decision-making is now the most common high-stakes use case for AI among UK adults. A quarter (26%) are using AI to figure out debt management strategies. Most of these conversations are happening inside ChatGPT or similar platforms, not inside the bank's own apps.

Source: Lloyds Banking Group, August 2025

Financial services is one of 7 sectors where 40 to 55% of consumers are already using AI search for purchasing decisions.

McKinsey's consumer-spending research identifies financial services alongside consumer electronics, grocery, travel, wellness, apparel and beauty as the top sectors where AI search has already become a primary purchase research channel. The discovery surface for financial products (savings, insurance, mortgages, investment platforms) is changing faster than most established players are tracking.

Source: McKinsey, New front door to the internet, November 2025

AI in fintech revenue is projected to grow from $15.4 billion in 2024 to $60.6 billion by 2033.

Straits Research puts the AI-in-fintech market at a 16.45% compound annual growth rate through to 2033. The growth includes both consumer-facing AI tools and the underlying AI infrastructure inside financial products. The product layer is moving in step with consumer adoption, not behind it.

Source: Straits Research, AI in Fintech Market Size, 2025

Does AI visibility work the same way on every platform?

This is the section that changed most in the August update. Until recently there was very little public data on whether showing up in one AI system meant anything for the others. Now there is, and the answer is no.

Of more than 1,200 brands tracked across 4 AI platforms, only 36 held top-100 visibility on all of them, every month.

Semrush analysed 126 million US AI search prompts between January and April 2026, covering ChatGPT, Gemini, Google AI Mode and AI Overviews across 22 industries. The 36 brands that turned up everywhere are the ones you would guess: YouTube, Google, Reddit, Amazon, Facebook, Apple, Walmart. Everything else is platform-dependent. A brand can be a fixture in ChatGPT answers and close to absent from Gemini. Visibility is not one number, and treating it as one number is how brands end up surprised.

Source: Semrush, AI Visibility Index 2026, June 2026

ChatGPT cites around 15 sources per answer. Gemini cites 3.

The same study found the platforms behave differently enough to need separate measurement. ChatGPT draws widely and leans on community sources. Gemini pulls from a much smaller pool, and the overlap between the brands it names and the websites it cites can be as low as 30%, which means it will recommend one company while quoting another. Being the source that gets cited and being the brand that gets recommended are two different jobs, and you can win one without the other.

Source: Semrush, AI Visibility Index 2026, June 2026

45% of marketing leaders cannot measure their brand's visibility in AI answers. Only 9% have the tools to track it properly.

Semrush's companion survey suggests the measurement gap is now wider than the strategy gap. The same survey found that teams running SEO and AI visibility as one workflow reported traffic or lead gains from AI platforms 81% of the time, against 36% for teams keeping them apart. That matches what I see in client work. The AI visibility problem is usually a content and entity problem wearing a new name.

Source: Semrush, AI Visibility Index 2026, June 2026

How much do people trust AI search answers, and should they?

These figures reframe what AI visibility is really about. Not traffic. Description.

When an AI Overview appears, users click a link inside it just 1% of the time.

Pew Research tracked the real browsing behaviour of 900 US adults across nearly 69,000 Google searches. When a search returned an AI summary, people clicked a source cited inside that summary in only 1% of visits. They clicked any traditional result in 8% of visits, against 15% when no summary appeared. People read the answer and move on. They rarely check where it came from. If a business is described incorrectly in that summary, almost nobody is clicking through to find the correct version.

Source: Pew Research Center, July 2025

People end their search session 26% of the time after seeing an AI summary, against 16% without one.

The same Pew study found that an AI summary is more likely to be the end of the road. Users stopped browsing entirely after 26% of pages that carried a summary, compared with 16% of pages with only traditional results. The answer at the top of the page is increasingly the whole interaction, which puts a lot of weight on that answer being right about you.

Source: Pew Research Center, July 2025

Google AI Overviews are accurate around 85 to 91% of the time, which still leaves hundreds of millions of wrong answers a day.

An analysis by AI startup Oumi for The New York Times ran more than 4,300 Google searches against the SimpleQA benchmark. Overviews were accurate about 85% of the time on the older Gemini 2 model and 91% on Gemini 3. At Google's scale, even a single-digit error rate means a very large number of confidently wrong answers reaching real people every day. Accuracy improving is not the same as accuracy solved.

Source: Oumi and The New York Times analysis, reported April 2026

What these AI search statistics mean for your brand

4 categories. 4 different paces. Same underlying shift.

Consumers in 2026 are not asking "is AI better than Google?" They are using both and forming opinions inside whichever surface gives them the clearest answer. The brands being recommended are the ones AI systems can confidently describe. The ones being skipped are the ones AI cannot make sense of, regardless of how strong their traditional SEO is.

Most marketing teams are still optimising for the wrong surface, and not because they doubt the data. They cannot see their own position in it. 45% of marketing leaders say they cannot measure how their brand shows up in AI answers at all, which makes the rest of the conversation hard to have.

Google added a control in June 2026 that lets site owners keep their pages out of AI Overviews and AI Mode. Almost nobody should use it. If your pages are not in the grounding set, the answer about you gets written from whatever else is lying around.

The work that closes the gap is not glamorous. Entity clarity. Schema. Consistent positioning across every surface AI systems can read. Citations from sources the models trust. A press boilerplate that says the same thing your About page says. None of it is new craft. It is the foundational content work, applied to a new audience.

The brands that act on this data are the ones who will be cited when an AI is asked who solves a problem in their category. The brands that wait are the ones being described by their competitors' content.

See how AI describes your brand right now. Everwilde One's free AI Search Entity Builder takes 3 minutes and gives you 3 outputs you can use today: an AI Entity Statement, a press boilerplate and a founder profile, all written in the language AI systems are using to describe your category. Start the builder at everwildeone.com/ai-search-entity-builder.

---

About this roundup. This page collects the most reliable AI search statistics we can find, and we update it every month. It was last updated in August 2026. Every statistic was published in 2025 or 2026 by a named research firm, platform, or reputable third-party publication. Sources are linked inline. Stats were verified against their primary source before publication. Compiled by Michelle Legge, founder of Everwilde One, a content-centric AI visibility agency based in Cape Town.