SourceNode research

What McKinsey’s AI search research says, and what to do about it

What McKinsey’s AI search research says, and what to do about it

What McKinsey’s AI search research says, and what to do about it

McKinsey found that a business’s own sites are only 5 to 10% of what AI search references, and that 44% of AI search users now call it their primary source. What the numbers mean and where they stop.

McKinsey found that a business’s own sites are only 5 to 10% of what AI search references, and that 44% of AI search users now call it their primary source. What the numbers mean and where they stop.

Odin Lund

Research, SourceNode

Primary source of insight for AI search users
AI-powered search44%
Traditional search31%
Retailer or brand websites9%
Review sites6%
Source: McKinsey AI Discovery Survey, August 2025, n=1,927
Primary source of insight for AI search users. Source: McKinsey AI Discovery Survey, August 2025, n=1,927
AI-powered search44%
Traditional search31%
Retailer or brand websites9%
Review sites6%

Key findings

  • 44% of AI search users call it their primary source of insight, against 31% for traditional search. McKinsey, 2025.

  • A business’s own sites are only 5 to 10% of the sources AI search references, in many cases. McKinsey, 2025.

  • AI tools were the biggest influence on the most recent purchase for 29% of AI shoppers, against 9% for search engines. McKinsey, 2026.

What did McKinsey find about AI search?

McKinsey published “New front door to the internet: Winning in the age of AI search” in October 2025, based on a survey of 1,927 US consumers in August 2025. Three findings carry most of the weight.

  • Half of consumers use it. Half of those polled now intentionally seek out AI-powered search engines.

  • It is already the first stop for its users. 44% of AI search users say it is their primary and preferred source of insight, ahead of traditional search at 31%, retailer or brand websites at 9% and review sites at 6%.

  • Most businesses are not measuring it. Just 16% systematically track their AI search performance.

McKinsey also projects that $750 billion in US revenue will pass through AI-powered search by 2028, and that businesses that do not prepare could lose 20 to 50% of their traffic from traditional search. Those two are forecasts, not measurements, and should be quoted that way.

Where do AI answers about a business come from?

Mostly from pages the business does not own. This is the finding that changes the work. McKinsey writes that in many cases a business’s own sites make up only 5 to 10% of the sources AI search references. In consumer packaged goods and financial services, more than 65% of the sources are publishers, user-generated content and affiliate sites.

SourceNode’s own research shows the same shape. In one legal market, across 4,222 AI answers in 30 days, the engines made 18,742 retrievals from 46 third-party domains against 4,319 from the leading business’s own site. Roughly four in five came from pages that business did not own.

Ahrefs found something similar from another direction. Across 75,000 businesses, mentions across the web correlated with visibility in Google AI Overviews at 0.664, against 0.218 for backlinks. That is correlation, not proof of cause, but it points at the same place: what other sites say about you.

What changed in the 2026 follow-up?

In October 2026 McKinsey published a second study, surveying about 21,000 consumers in 11 markets. It moves from search to shopping, and the numbers are larger.

  • 53% used AI for shopping in the past three months, on average across the markets surveyed. The US figure was 48% and China’s was 80%.

  • AI now outweighs search engines for its users. AI users were 3.2 times more likely to rank AI search tools as more influential than traditional search engines.

  • It decides purchases. AI chat or search tools were the single biggest influence on the most recent purchase for 29% of AI shoppers, against 9% for search engines.

The base for the last two figures is people who already shop with AI, not all consumers.

What to do about it, and sources

Read the figures with their limits. One line in the 2025 article says about half of Google searches already show AI summaries. McKinsey attributes that to trend analysis. Measured studies are lower and vary by keyword set: Pew found an AI summary on 18% of Google searches in March 2025. Name the tracker whenever you quote a prevalence figure.

Then act on the part that is solid. Find out what the engines say about you today, by engine. List the third-party pages they read to say it. Correct those, publish against the questions where nobody is named, and keep measuring. That is the work SourceNode does, described on our product page, and it is why we are data and research first. How we measure results.

The researcher

Odin Lund is co-founder and COO of SourceNode and leads its data analytics and research, including the daily tracking behind this work across ChatGPT, Gemini and Google AI Overviews. Odin founded First Query, the AI search practice that became SourceNode, and co-founded The Pixel Paper.

Meet the research team

About this research

Published by SourceNode

SourceNode is a data and research-first AI search firm. We record what AI models say about a business every day, trace each claim to its source, and use that data to get businesses named in AI answers. Every figure we publish carries its method and its sample.

Share Blog