McKinsey's State of the Consumer 2026 report, published June 22, 2026, tells brands to invest in public relations because AI shopping tools draw mostly on sources brands do not own. McKinsey, citing XEO360 data from October 2025 to May 2026, found that brand-owned websites supply only 1 percent of the sources large language models cite for consumer goods. Forums, retailer reviews, blogs and video platforms are among the sources AI tools pull from.
In this article, a large language model (LLM) is the software behind tools such as ChatGPT, Gemini and Perplexity, and a citation is a source that an AI answer names or links. Generative engine optimization (GEO) is the work of making a brand's pages and mentions easy for those tools to find and quote. ConsumerWise is McKinsey's recurring US consumer survey.
What does McKinsey's report say about AI shopping?
McKinsey's State of the Consumer 2026 report names the tech-driven path to purchase as one of four trends reshaping consumer behavior, alongside the health revolution, the experience economy and the resourceful consumer. The report's five authors, including Anna Pione, a partner in McKinsey's New York office, represent the firm's Consumer Packaged Goods and Retail Practices.
The report draws on a survey of 4,863 consumers in Brazil, France, Germany, the United Kingdom and the United States, taken from March 31 to April 8, 2026. In that survey, 28 percent of Gen Z respondents said they already use generative AI tools for shopping, compared with 16 percent of baby boomers. Sixty percent of Gen Z said they regularly use the AI overview at the top of search results, against 29 percent of boomers.
Search behavior is shifting alongside that adoption. McKinsey reports that open web traffic has fallen 8 percent since 2023, and that shoppers increasingly find answers inside AI-generated results and click through to fewer brand, retailer and publisher sites.
The firm's US ConsumerWise survey adds detail. Published March 2, 2026, it found that 68 percent of respondents had used at least one AI tool in the previous three months, and that 44 percent of AI search users named it their most preferred source of information, ahead of search engines at 31 percent, retailer and brand sites at 9 percent and review sites at 6 percent.
By late July and early August 2026, 46 percent of US respondents said they would definitely or probably use AI tools for holiday shopping, and 49 percent of likely users planned to compare products or prices, according to McKinsey's August 27, 2026 ConsumerWise update.
Where do AI shopping tools get their information?
AI shopping tools get most of their product information from sources outside brand websites, according to McKinsey's analysis of which sources large language models cite. Even among the ten most frequently cited websites, which together account for about 22 percent of all citations, brand websites make up at most 10 percent, based on XEO360 data covering October 2025 to May 2026.
Affiliate blogs carry even more weight in one dataset. McKinsey's ConsumerWise analysis of data from Semrush, a digital marketing platform, shows that affiliate blogs, sites that earn commissions by promoting products through tracked links, supply 50 percent of citations in AI-generated search summaries. That equals the combined share of user-generated content, academic and market research, brand and retailer sites, and news and media.
McKinsey describes the process as aggregation. When a generative AI platform answers a query, it combines information from many third-party sources, which makes it harder for a brand to control how it is presented. If those sources disagree or leave gaps, McKinsey says, AI tools are less likely to include the brand or describe it accurately.
What does McKinsey tell brands to do about it?
McKinsey tells brands to shape the first-party and third-party content that large language models draw on, and it names public relations as part of that work. The report says brands should strengthen public relations, keep messaging accurate and consistent across platforms, and keep an active presence in the forums where consumers already discuss them.
Kevin McCauley, a reporter at the PR trade publication O'Dwyer's, summarized the finding on June 25, 2026, in O'Dwyer's coverage of the McKinsey report, headlined "McKinsey Tells Brands to Boost PR." McKinsey adds that earned media and third-party validation become more critical as a result.
The report also recommends changes on a brand's own site: clear, structured, scannable pages such as FAQs and lists, and detailed material that was not public before, such as clinical research or technical documentation.
The August 27, 2026 update adds a warning about prices. McKinsey says AI tools that draw on frequently updated data can surface outdated or conflicting prices, which could deter shoppers from going further.
My read is that the report validates earned media. PR teams have spent two decades measured against paid channels with cleaner attribution, and McKinsey's data now shows that most of the pages AI tools cite are pages other people wrote. An earlier post on this site about Gartner's forecast for PR budgets covers the budget side of the same shift.
How can a brand turn that advice into a plan?
A brand can turn McKinsey's advice into a five-step plan that starts with measuring what AI tools say today and ends with repeating that measurement every quarter. The table below pairs each step with the reason it works.
| Step | Action | Why it works |
|---|---|---|
| 1. Audit | Ask ChatGPT, Gemini, Perplexity and Google AI Overviews the same ten category questions each month. Log whether the brand appears, how it is described and which sources are shown. | The log shows which outlets, forums and review pages the tools pull from in the brand's category, so later work targets those sources. |
| 2. Reconcile | Write one fact sheet covering founding date, leadership, product claims and prices. Correct conflicts across the brand site, retailer pages, profiles and press kits. | McKinsey says inconsistent or incomplete information makes AI tools less likely to include a brand or describe it accurately. |
| 3. Earn | Pitch the outlets, reviewers and creators already cited in the audit, with named experts and original data. | Brand websites supply about 1 percent of citations in consumer goods, so coverage on cited third-party pages is how a brand's facts reach the answer. |
| 4. Participate | Have identified employees answer real questions in the forums where the category is discussed, and state who they work for. | McKinsey lists forums among the third-party sources AI tools combine. Named, disclosed answers add accurate facts to those threads. |
| 5. Publish and re-measure | Add FAQ pages, list pages and technical documentation to the brand site, then repeat the audit every quarter. | McKinsey recommends structured, scannable pages and detailed material not previously public. The quarterly repeat shows whether new coverage changed what the tools say. |
Illustrative scenario: a skincare brand asks Perplexity who leads its category and sees a competitor's review roundup cited in the answer. That brand now knows which publication to pitch with its own named experts and data, and which facts on its retailer pages need correcting first.
Step 1 should log two numbers, how often the brand appears and how often its own pages or earned coverage are cited, because those rates can move independently. A post on this site about fintech AI visibility measurement reports that appearance rate and citation rate move independently, and the AI visibility research library collects related findings by industry.
Step 2 also covers corrections. An outdated claim on a retailer page or a forum thread can remain available to AI tools until someone updates it, so the fact sheet should list every page that carries a price, a specification or a leadership name. Brands with many local units face this at scale, and a post on franchise AI visibility explains why AI engines can fail to treat thousands of units as one brand.
Step 3 depends on credible spokespeople. The executive communications guide on this site covers how to build an executive's voice with reporters and AI engines. For growth teams, growth marketing and AI search covers how acquisition work changes when shoppers start with an AI tool.
The plan does not work through volume. Mass-produced releases and thin pages add weak sources to the pool AI tools draw from, and McKinsey's advice centers on accurate, consistent and credible material, not on more of it.
What are the limits of the McKinsey data?
McKinsey's data has three limits that brands should weigh before shifting budget: the citation figures cover consumer goods, shoppers still use AI mainly for early research, and traditional search has not disappeared.
The 1 percent brand-website figure comes from consumer goods queries in XEO360 data, so a software, financial or healthcare brand may see a different split. Brands outside consumer goods should run the step 1 audit before assuming the same pattern.
The ConsumerWise survey found that AI-enabled shoppers use the tools mainly to compare options and learn about a category, and fewer use them to build a basket or check out. McKinsey says that is the pattern for now, which means the checkout stage may change.
The 8 percent drop in open web traffic since 2023 does not prove that AI caused it. McKinsey says traditional search remains highly relevant and total search activity continues to grow, even if the number of searches per user declines. Gen Z respondents also told McKinsey they are skeptical of both social media and generative AI, despite using both widely.
The takeaway for brand and PR leaders
The takeaway from McKinsey's report is that a brand's reputation now reaches shoppers through pages the brand does not control. Brands that keep their facts consistent and earn coverage on the sources AI tools already cite give those tools accurate material to use. Brands that do neither leave the answer to whoever wrote the pages the tools find.
Last updated: October 13, 2026.