Instacart Brings Agentic Grocery Shopping to Gemini

Instacart becomes the first grocery partner integrated into Gemini, powering AI-driven grocery commerce

Today, we're announcing that Instacart is now integrated with Gemini, Google's AI assistant, making us the first grocery partner on the Gemini platform. Users can connect their Instacart account within Gemini and build a real, shoppable grocery cart through natural conversation, then head to Instacart to place their order for delivery or pickup. The reason this works comes down to Instacart’s intelligence and what makes grocery different.

Most commerce categories are event-driven. Grocery is habitual: the same brands, the same dietary constraints, the same preferred substitutions, week after week. That behavioral regularity is actually what makes grocery one of the strongest use cases for agentic AI. The patterns are consistent and learnable, which gives the system something real to build on. But it also raises the bar considerably. A query-response model isn't enough here. The system has to know the customer – their history, their preferences, their household – not just parse the request. That's a meaningfully harder problem, and frankly, most companies don’t have the data infrastructure to solve it. 

But Instacart does. 

For years, Instacart has been building the intelligence layer for grocery – the data infrastructure, personalization, localization, and fulfillment connectivity that transforms everyday intent into groceries delivered. When a customer links their Instacart account through Gemini, they unlock personalized results based on how they shop: their usual brands, dietary needs, preferred substitutions, and local store. The more you shop with Instacart, the smarter our intelligence layer gets for you, surfacing the right products, flagging the right replacements, and helping you discover things that fit your household's needs. 

We built this infrastructure to meet customers wherever the grocery journey begins – with a recipe, a weekly meal plan, a pantry restock, or for some, in a conversation with an AI assistant. Whatever the starting point, Instacart's intelligence layer is what makes the grocery experience actually work, translating natural language input into a real, shoppable cart that is accurate, local, and built around how you shop. 

Connecting to Instacart while using Gemini is simple. Gemini users can start with a prompt like, "Add everything I need for a lemon-dill salmon recipe to my Instacart cart," or "Build me a weeknight dinner plan on Instacart for my family of three." Gemini recognizes the intent, prompts the user to connect their Instacart account, and from there, they select their preferred retailer and go straight to building a cart. Instacart's catalog and personalization infrastructure does the matching in real time and items land directly in a live cart, synced with the user's Instacart account throughout the session, so nothing gets lost along the way. Gemini renders the experience through Google's own interface, populated with the real-time data Instacart surfaces: live inventory and product recommendations shaped by that user's history and preferences.

As the conversation evolves – swapping one ingredient for another or adjusting quantities – the cart updates accordingly, all through the use of natural language. When the cart is ready, a single tap takes the user to Instacart to complete their order. 

Grocery is one of the hardest categories to execute reliably in AI. When a user says "add everything to my cart for a taco night dinner for four," the system has to handle ingredient recognition, gather item availability and price from nearby stores, and deliver relevant personalized results based on what that particular household actually buys – and make it feel seamless.

Instacart's catalog spans more than 2,200 retailer banners and nearly 100,000 stores across North America, with 22 million unique items behind every query, giving the system the breadth needed to match a conversational request to an actual, purchasable product at a user's local store. Our rapid inventory data means recommendations reflect what's on shelves today, not a cached snapshot from earlier in the week. Smart Shop, our AI-driven personalization experience, draws on purchase history and selected dietary preferences to ensure the cart feels built for the customer’s specific household, not a generic one.

Gemini is the latest in a series of major AI platforms, including Claude and ChatGPT, where Instacart has launched an agentic grocery integration. Across each, Instacart's catalog, real-time inventory, personalization, and fulfillment network are made accessible through the chat interface a user is already in. That same intelligence layer powers grocery experiences for retailers, and on our own marketplace. The consistency across all of these is by design: we’ve built the grocery infrastructure and intelligence for the agentic AI era that works everywhere customers shop, rather than a series of one-off integrations.

Building native grocery experiences across today's leading AI assistants and earning the position of first grocery partner on Gemini reflects the engineering investment behind our platform approach – building the grocery layer once, not retrofitting integrations after the fact. The catalog, inventory data, personalization, and fulfillment network are the same across every surface we've launched on. That consistency is what makes this scalable. As AI assistants take on more of the everyday task layer, grocery is one of the most frequent, most complex categories they'll need to get right. We've built the infrastructure to support that and are ready to support whatever interface comes next.