Instacart’s AI Now Builds the Whole Cart. It’s Also Trained to Find Cheaper Alternatives.
Instacart launched Clementine on September 9, an AI assistant that turns a message like “high-protein easy dinners for two” into a ready-to-buy cart. It's live across North America for most US and Canadian customers, built on 1.6 billion lifetime orders and 10 million daily inventory signals.
For CPG brands, the capability worth reading twice is this one: Clementine surfaces deals and lower-cost alternatives as a core function. The assistant isn't just filling a cart. It's actively looking for cheaper substitutes.
What Clementine Actually Does
The assistant generates recipes, recommends products based on household preferences, reorders frequently purchased items, and builds lists for specific occasions. A prompt like “everything for a football watch party for 20 people” returns a tailored cart accounting for brand preferences and dietary restrictions.
It also handles list uploads. A shopper can photograph a handwritten list or screenshot a digital one, and Clementine converts it into products added to cart.
Because it draws on Instacart's order history, catalog, and live inventory, it knows what's on the shelf at the shopper's chosen store, what's on sale there, and what that household buys week after week.
The Enterprise Half Matters More for Distribution
Alongside Clementine, Instacart expanded Cart Assistant, its white-label version that runs on retailers' own websites and apps under their branding. Food Bazaar, Heritage Grocers Group, and Woodman's are live now. ALDI US, Harmons, The Save Mart Companies, and Stew Leonard's are among those expected to follow.
That's the part that changes scale. Cart Assistant combines Instacart's grocery intelligence with the retailer's own catalog and customer data, letting grocers customize name, tone, loyalty integration, and recipe content. The same selection logic ends up operating across both Instacart's marketplace and a growing number of individual grocery sites.
Instacart Isn't First, and That's the Real Story
Kroger launched a chainwide assistant in January. Albertsons says its own tool cuts a 46-minute shopping trip to four minutes. Instacart is joining a category, not creating one.
As PYMNTS framed it, the competitive fight isn't over who builds the smartest AI. It's over whose inventory data is accurate enough to trust, since a recommendation is worthless if the customer arrives and the product is gone. That's where Instacart's 10 million daily inventory signals become the actual asset.
About the Basket Size Claim
Instacart says orders placed with Clementine contain more items on average and exceed the company's $115 average basket size. Early usage also shows customers starting with restocking before moving into recipe discovery.
Worth being precise: that's company-reported observational data from a product that launched two days ago. Larger baskets could reflect the assistant driving more spending, or they could reflect the kind of shopper who tries a new AI tool already being a heavier buyer. The data doesn't separate those.
What This Means If You Sell CPG or Grocery Products
The substitution function is the thing to understand. An assistant instructed to surface lower-cost alternatives will, by design, sometimes route a shopper away from your product toward a cheaper one, and it makes that decision before the shopper ever browses a category page or sees your packaging.
That shifts where competition happens. Price positioning relative to alternatives in your category matters more when an algorithm evaluates it automatically on every cart build. So does purchase history, since “order my usuals” favors whatever the household already bought, making existing customer relationships more defensible and new customer acquisition harder.
The same dynamic is showing up elsewhere. Amazon's Alexa for Shopping recommends products that often don't match search rankings, meaning AI selection runs on different signals than the ones sellers have optimized for. In grocery specifically, the signals appear to be availability, price relative to alternatives, and purchase history rather than listing optimization.
Whether paid placement factors into AI-built carts is the open question Instacart hasn't addressed. Given that Instacart runs a substantial advertising business, that's worth watching closely, since the answer determines whether CPG brands can influence cart inclusion at all or simply have to compete on price and availability.

