Key takeaways
• Traditional search is not in retreat: only 13% of consumers start a major purchase journey with an AI tool, against 36% who start with a search engine. Google’s search query volume hit an all-time high in Q1 2026.
• AI extends the research journey but does not close it. Only 8% of consumers who started with AI felt comfortable buying on that recommendation alone. Overall, consumer trust in AI recommendations stands at just 15%.
• Category matters. AI adoption in retail is moving fastest for rational, utilitarian purchases. For fashion and luxury, where desire and personal taste drive the decision, trust and conversion lag considerably behind.
Will AI replace search engines? Based on the data available in mid-2026, the answer — at least in the context of retail and shopping — is no.
For two years, the dominant narrative has run in one direction: traditional search is dying, Google is losing ground, and AI tools like Perplexity, ChatGPT and Gemini are becoming the new front door to every consumer decision. It is a compelling story. It is also, so far, only partially true.
Research presented at Fevad’s Grande Conférence in Paris in July 2026 by the Chaire TREND(S) at Toulouse School of Management — the latest study to test AI and traditional search head-to-head across the full purchase journey — found that AI tools underperformed traditional search at every stage, across 900 consumers in France and the US. Shoppers using AI asked more questions and were significantly less likely to complete a purchase. The finding is consistent with a growing body of research from Razorfish, Alchemer, McKinsey and Bain, all pointing toward the same conclusion: AI has changed how consumers research. It has not yet changed how they buy.
Are consumers really switching to AI search?
The data points in one direction. In a May 2026 study, Razorfish surveyed consumers who had recently made a major purchase. Thirty-six percent said they started their journey with a search engine; only 13% started with an AI driven search tool. It is a significant gap, and nothing in the data available as of mid-2026 suggests it is closing.
Google’s own Q1 2026 earnings told much the same story: search revenue up 19% year-on-year, query volume at an all-time high. Notably, that growth is happening alongside the rise of AI tools. Hundreds of millions of people using ChatGPT, Gemini and Perplexity every week, not in spite of them.
A separate Bain & Company survey of 1,500 US consumers, conducted in late 2025, found that 44% of online buyers mostly start their journey in an AI tool or split their time between AI and traditional search engines. That figure sounds like a larger shift than it is: Bain includes in its count those who use both tools interchangeably, which inflates the apparent movement away from search. The Razorfish numbers, which isolate where consumers start, are a cleaner read on primary behavior.
Consumers are not abandoning search engines. They are adding AI on top. Both tools are running in parallel. At least for now.
What do consumers ultimately do when AI gives them an answer?
They keep searching. That is, broadly, what the data shows. In fact, among consumers who used AI as their first information source in a purchase journey, only 8% said they were comfortable buying immediately on the basis of that recommendation, according to Razorfish. The other 92% wanted more: price comparisons, reviews from real buyers, confirmation from the brand’s own website. Far from collapsing the purchase funnel, AI appears to be extending it, adding a layer of research before consumers turn to the tools they already trust to make the final call.
Trust is the underlying problem. According to the Alchemer 2026 Retail Report, which surveyed 1,002 US shoppers, only 35.4% trust AI recommendations completely or mostly, and 22.2% do not trust them at all. When asked which source they trust most when making a purchase decision, online reviews came first at 36%, followed by friends and family at 31%. AI recommendations ranked third, at 15% — ahead of expert review sites and influencers, but a long way behind the sources consumers have relied on for years.
Razorfish puts it plainly: 78% of consumers are comfortable letting AI narrow down their options but still want to make the final choice themselves. Another 69% specifically want AI involved in research while retaining full control over the purchase decision. The technology is useful, in other words, up to a point. That point is somewhere before checkout.
As the Chaire TREND(S) at Toulouse School of Management found, AI search tools appear to trap shoppers in information loops, generating more questions rather than driving toward a decision.
Will AI replace search engines for every type of purchase?
The short answer would be yes, but the picture is way more nuanced. And the gap between product categories is significant.
Strategy&’s Fashion Retail Outlook 2026 found that AI agents are being adopted first for shoes, sportswear, denim, outerwear and basic items, categories where the purchase is largely rational, driven by utility and price comparison rather than emotional reasoning. Only 26% of consumers in their survey felt confident that an AI assistant could accurately match their personal style. For the categories where style, aspiration and desire drive the decision, consumer confidence in AI drops considerably.
Luxury sits on the more complicated end of that spectrum. According to the BCG x Altagamma True-Luxury Global Consumer Insights 2026, based on 10,000 luxury consumers across 11 markets, 79% of luxury consumers already use AI to research products, asking for recommendations and comparing options across leather goods, watches and hospitality. AI/GenAI has earned a net trust score of 29 percentage points among luxury consumers, ranking fourth among main information sources, on par with traditional web search and not far behind word of mouth at 36 points. Only the brands’ own websites ranks higher, at 42 points.
The BCG data tells a story of genuine adoption: luxury consumers are AI-literate, they use it regularly, and they trust it more than the industry anticipated. What it does not show is conversion.
What does this mean for brands?
The question, “will AI replace search engines altogether”, may matter less to brands than where their customers are making decisions right now, and how fast that is changing.
McKinsey’s State of the Consumer 2026 found a sharp generational split. Twenty-eight percent of Gen Z use AI tools for shopping, compared to 16% of baby boomers. But even among the youngest consumers, net trust in AI sits at around 20%, against more than 50% for traditional search engines. Adoption and trust are moving at different speeds, and for brands, that gap matters.
Bain’s research points to something more structural. Eighty-nine percent of unbranded AI prompts are fulfilled by third-party sources — reviews, editorial coverage, earned media. Brand-owned content barely registers. In an environment where AI increasingly shapes the first impression, the brands most likely to appear are those with the strongest independent footprint across trusted external sources.
Razorfish adds a forward-looking note: when consumers were asked where they would prefer to start purchase journeys in two years, 43% said search engines and 29% said AI. The shift is real. It is also slower than most of the current conversation implies.
Classic search optimization remains necessary, but it is not sufficient on its own. Most consumers are now moving between tools: starting with AI, verifying with search, checking a review site before committing. The journey has more steps in it, not fewer.
Search revenues at a record high. Consumer trust in AI recommendations at 15%. Right now, the safest assumption is that your customer is using both. And trusting neither completely.
