← The Growth Lab · AGENTIC COMMERCE
When AI Shops: Competing to Be Selected, Not Just Seen
Oct 26, 2026 · Read time: 3 min · DV8 Pacific
Today, AI assistants help shoppers decide. The next step is already under way: assistants that act on the shopper's behalf. They compare options, build a basket and complete checkout. This is agentic commerce, and it changes the basic rules of competing for a sale.
From attention to selection
Traditional advertising competes for attention: make the shopper notice your brand. Retail media competes for visibility and conversion on the digital shelf. Agentic commerce competes for selection: make the AI choose your product.
The difference is stark. Ask an assistant for "everything I need to make beef chow fun tonight" and it does not return a page of oyster sauces. It picks one. The product it selects gets the sale. The others are never seen. Ranking for the keyword "oyster sauce" matters less than being the product the AI chooses when it fulfills meals, occasions and shopping missions.
A new kind of digital shelf
Digital shelf work today means making sure a product looks good and ranks well on Amazon, Walmart and other retailers. In agentic commerce, it means making sure the AI understands when your product is the right choice for a shopper's need. That calls for a new measure alongside share of search: agent selection share, or how often an AI picks your product when completing relevant shopping tasks.
Brand still matters
It would be easy to conclude that branding stops mattering when AI makes the choice. It does not. Shoppers still say "I only buy Kikkoman," and an agent will follow that instruction. Shoppers ask for "a premium soy sauce for a gift," and brand positioning shapes what the AI considers premium. Trust, reputation and preference all feed into the selection.
What changes is that brand alone is no longer enough. The AI must also be able to read, understand and justify the product. That puts a new set of assets on the same level as advertising: machine-readable product data, complete attributes, accurate pricing, reliable availability and a credible base of reviews.
How large could this become?
Published forecasts from banks, consultancies and research firms vary widely, mostly because each defines agentic commerce differently. They agree on direction: a meaningful share of US online spending is expected to pass through AI agents by the end of the decade.
Which categories move first
The value of an AI agent grows with the difficulty of the decision. Categories with many choices, a need for explanation, personalization, higher prices or higher risk benefit most: skincare, small appliances, premium food and gifts. For everyday groceries, the prize is different but just as real: being the product chosen when an AI builds a basket for a recipe or a weekly shop.
Brands do not need to wait for agentic commerce to arrive fully. The groundwork is the same work that improves AI visibility and retail conversion today: clear, complete, structured product information that an AI can trust.
Sources: Published agentic commerce forecasts from Morgan Stanley, Bain, McKinsey and Gartner, as summarized by MetaRouter and commercetools, 2026; DV8 Pacific analysis.