Instacart Tests AI Assistant Ad Formats as Cart Assistant Expands

Instacart is testing three advertising formats built for its AI-powered Cart Assistant, aiming to convert conversational shopping into a new revenue stream. Sponsored recipes and occasions, deal surfacing, and product exploration ads will let brands insert themselves into planning conversations, Ali Miller, Instacart's general manager of advertising, told Marketing Dive. Testing is scheduled for the second half of 2026, after Cart Assistant rolls out across the company's full North American marketplace, Adweek reported.

The new formats extend an advertising business that already generates more than $1 billion in annual revenue across nearly 100,000 stores and 2,200 retail banners, Instacart said in a May press release. Existing sponsored content already pulls shoppers toward brands they were not searching for. Recipe ads deliver 78% out-of-aisle impressions and 43% new-to-brand sales, and occasion ads drive 90% out-of-aisle impressions and 36% new-to-brand sales, the company said. Cart Assistant itself is built on more than 1.6 billion lifetime orders and live inventory data. Orders placed through the AI tool are already larger on average than Instacart's $115 average order value, according to its second-quarter earnings release.

The push into conversational commerce comes as every major AI platform builds shopping into its own products. Google launched Universal Cart at its I/O conference in May, a single basket spanning Search, Gemini, YouTube and Gmail, with initial checkout including Target, Walmart, Sephora and Ulta Beauty, Stellagent reported. Ulta Beauty, which carries an Alpha Score of 48/100 in AlphaScala's proprietary stock analysis, is among the retailers participating. Amazon consolidated its own AI shopping tools into a single assistant called Alexa for Shopping, PYMNTS reported. Google separately expanded AI Mode search to let users connect with apps including Instacart, TechCrunch reported.

Instacart's advantage is grocery-specific data and fulfillment infrastructure that sits under the AI. Grocery is habitual rather than event-driven, with the same brands, dietary constraints and preferred substitutions repeating week after week, Instacart said. That gives the system consistent patterns to learn from. The advertising ecosystem already spans a consumer marketplace, an enterprise platform and the advertising layer itself, each segment feeding growth in the others. Cart Assistant adds a surface built around planning and discovery rather than transaction volume alone. Because the tool moves a shopper from a single-item search toward planning an entire meal, it naturally surfaces more products per conversation than a search bar built to answer one query at a time.

The test formats will appear after Cart Assistant expands to Instacart's full North American marketplace. On May 19, Instacart became the first grocery partner integrated with Google's Gemini, letting users connect their account and build a shoppable cart through natural conversation. That timing ties the ad bet to a larger distribution push beyond Instacart's own app.

The risk for Instacart is execution and competition. Rivals like Amazon and Google are building their own conversational shopping experiences, and retailers may develop in-house alternatives. If Cart Assistant fails to gain adoption or if the new ad formats do not convert at rates comparable to existing placements, the revenue uplift may fall short of expectations. The company is betting that the conversational layer will increase basket size and that advertisers will pay for placement within planning conversations. The 2026 test results will provide the first real signal.