
Shoppers who use AI are 3.2 times more likely to rank AI search tools above traditional search engines as an influence on what they buy. The finding comes from a McKinsey survey of about 21,000 consumers in 11 markets, published on 28 October. On average, 53 percent of the consumers surveyed already use AI at some point when they shop.
What happened. McKinsey asked consumers in June 2026 how they use AI chat and search tools before, during and after a purchase. Nearly a third of AI shoppers named these tools as the single biggest influence on their most recent purchase.
Why it matters. The first comparison of options increasingly takes place inside an AI answer. A brand that is missing from that answer is not merely ranked lower. For that buyer, it is not part of the market.
What you can do. Establish whether you are named when buyers ask about your category, make sure the sources AI services draw on describe you accurately, and track the result over time rather than testing once.
What the survey shows
The survey spans Australia, Brazil, China, France, Germany, India, Japan, Mexico, Saudi Arabia, the UK and the US. Its findings point the same way: AI tools are becoming the place where purchases begin, not a curiosity at the margins.
Research starts in AI. AI users are almost twice as likely to begin their research with AI search as with a traditional search engine or another online source.
AI search outweighs other influences. Beyond its lead over traditional search, AI users are 2.6 times more likely to rank AI search above other strong influences, such as friends and family.
General assistants lead. 56 percent had used a stand-alone AI chat or search tool when shopping, against 27 percent for a retailer’s own assistant. Outside China, ChatGPT and Gemini were the most used.
Transactions are following. In the US, the share of AI users who completed a purchase through AI rose from 10 to 17 percent between December 2025 and June 2026.
The reasons consumers give are practical. 81 percent say AI is good at narrowing their choices, and 79 percent say it saves them significant time. Across the markets, an average of 73 percent of AI users say they trust AI chat and search tools.
What the data does not show
Two limits matter. The survey covers consumers rather than business buyers, and no Nordic market is included. The figures describe how people say they behave, and are best read as a direction rather than a forecast.
The closest European reference is an earlier McKinsey study of 749 respondents in France, Germany and the UK, conducted in December 2025. There, 38 percent said they use AI tools to research products or services or to decide what to buy. Trust was highest when AI compares options and recommends a best choice.
Our view: the shortlist is the new first page
McKinsey closes with six priorities for companies. The first is to win the AI consideration set: to maximise the likelihood that a brand makes an AI agent’s shortlist. We think it is the most consequential point in the report, for three reasons.
From ranking to inclusion. Search rewarded position, and a company in fifth place still earned some clicks. An AI answer offers a recommendation instead of ten blue links. The question is no longer how high you rank, but whether you are named at all.
The decisive work happens beyond your own website. McKinsey puts third-party authority on the same list as machine-readable content. AI services form their picture of a company from many sources, not only its own pages. A strong website is necessary, but rarely sufficient.
Business buying will follow. The survey measures consumers, but the mechanism carries over. When a procurement lead asks an AI service to compare suppliers, the answer becomes the first shortlist. Companies left out will often not know what they have lost, because a buyer who never saw them leaves no trace.
McKinsey also argues that measurement has to be continuous, since AI agents do not pause. We agree, and would add that a single test can easily mislead. AI answers vary between runs, between services and over time, so the picture only becomes reliable when the same questions are tracked systematically.
What it means for you
The direction is clear, even where local data is still missing. The risk is not that AI sends buyers to a competitor by mistake. It is that buyers never learn you were an option.
Ask what your buyers ask. Put the questions your customers ask into ChatGPT, Claude, Copilot, Gemini and Google AI Overviews, and note who is recommended.
Map the sources. Identify which websites, databases and publications the answers draw on, and whether you are represented there accurately.
Close the gaps. Publish clear, factual pages that answer the questions buyers raise before they decide.
Measure over time. Track the same questions regularly. One snapshot says little about where you stand.
Mynth helps companies get named, cited and recommended when customers ask AI services, so that more of the buyers who start with AI find their way to you. If you want to know whether you make the shortlist today, talk to us.
Sources
Europe’s agentic commerce moment: Decision influence is here; execution is coming – McKinsey & Company
Who wins when AI does the shopping? New evidence on agentic commerce – McKinsey & Company, 28 October 2026