51% of US consumers used at least one AI tool to help with shopping in the past month, according to NIQ. The figure comes from the firm's Agentic Commerce Tracker and marks the first time the share has passed 50%. Use has risen steadily over the past two quarters and now spans the full purchase journey, from product discovery to the buying decision. The shift matters for business because AI assistance has moved from early-adopter behavior to a mainstream habit.
How AI spread across the purchase journey
Product recommendations were the most common AI shopping use, cited by 20% of consumers. AI personal shopping assistants followed at 16%. Shoppers mainly turn to these tools to compare options, assess value and narrow down choices. In NIQ's description, AI is "compressing decision-making" rather than replacing shopping itself. That pattern explains why adoption now covers every stage instead of a single task such as search or checkout.
The numbers come from a monthly survey of about 500 US consumers, conducted as part of NIQ's Quick Question research. NIQ said the tracker also covers Canada, although only US figures were published in this release. Separately, NIQ is building commercial measurement for the same shift. On 2 September it announced a joint product with Similarweb that will show how brands appear in ChatGPT, Gemini, Google AI Mode, Perplexity and Claude, alongside the traffic and sales those tools generate.
NIQ describes the development as a move from a search-driven path to purchase to an AI-driven one. President for North America Liz Buchanan said the change is happening faster than many organizations realize, adding that AI-assisted shopping was viewed as emerging behavior just months ago. For brands and retailers, product content, ease of AI discovery and structured product data now matter as much as shelf placement and search rankings. The background is fragmented: Similarweb data shows AI search spread across several assistants, while Visa, Mastercard and Ant International are working on identity checks for buying agents and grocery buying is already affected.
What this means for sellers and brands
For companies that sell goods, product information becomes a sales interface for machines as well as people. Complete specifications, consistent attributes, clear pricing and availability, and content that AI systems can parse decide whether a product is recommended or filtered out. Large retailers with established feeds and content operations can extend existing catalog processes to meet this requirement. Smaller sellers face a narrower starting point: clean up core listings first, because gaps in basic data now affect visibility in both search results and AI answers.
The 51% figure records how many people used an AI tool at least once, not the share of spending that flows through AI or the frequency of use. A monthly sample of about 500 respondents signals direction without breaking down categories, demographics or repeat behavior. Companies evaluating measurement tools should ask how a provider separates AI-assisted browsing from traffic and sales sourced from specific assistants. The joint NIQ and Similarweb product, due in the fourth quarter of 2026 for only a few categories and markets, will initially answer that question for a narrow slice.
A concrete marker to watch is whether the share holds above 50% in coming monthly tracker releases and whether the measurement product launches on schedule. Sustained readings plus working attribution from major assistants would confirm that AI discovery converts into measurable traffic and sales. If that link appears, assortment and content management will need the same operational attention as shelf and search.
