
Google is testing a new way to shop in India that could make the distance between a recommendation and a purchase much shorter. Some users can reportedly buy selected Flipkart products from within Gemini and AI Mode, with a Buy button leading into a Flipkart-branded checkout flow. It is a limited experiment, not a nationwide launch.
The test was reported by TechCrunch, which says it saw the experience and spoke with people familiar with it. The option is available to a small group of shoppers and a narrow range of products, including phones, electronics and mobile accessories. Google has not published a detailed announcement setting out eligibility, a firm launch date or the commercial terms of the arrangement.
That limitation matters. Anyone opening Gemini today should not assume they can order everything they see, and the feature does not mean an AI assistant is independently making purchases on a user’s behalf. The reported flow still asks a person to choose a product and proceed through checkout. What changes is where that journey begins and how few steps separate the answer from the transaction.
Google had already moved beyond a list of blue links. In an April update for India, the company said Gemini could show product listings, prices and comparisons while AI Mode could organise shopping answers around reviews, inventory and images. A checkout connection is the next commercial step; it lets a shopper act on a recommendation while the decision is still fresh.
For Flipkart, which is owned by Walmart, the appeal is clear. Its products can appear where shoppers are already asking questions rather than waiting for them to open a marketplace app. For Google, the test asks a bigger question. If search and an AI assistant become places where people complete purchases, Google is no longer only directing traffic to merchants. It is becoming part of the storefront.
That raises practical questions about who controls the customer relationship, how products are ranked and whether smaller sellers receive a fair chance to be seen. It also means shoppers will need to pay attention to the final price, delivery terms and returns policy, not just the AI-generated summary that brought them to the listing. A convenient checkout is useful only if the purchase details remain transparent.
Google has also discussed wider tools for AI-assisted commerce, but the Flipkart experiment appears to use a retailer-branded checkout rather than the Google-hosted experience demonstrated elsewhere. It is important not to treat every agentic-shopping feature as the same product. Different checkout designs can give the platform and the retailer very different roles.
The competitive pressure is already visible. Meta has tested shopping recommendations inside its own AI assistant, while Google is trying to turn discovery into action. The Flipkart trial gives that ambition a concrete setting in one of the world’s largest online retail markets. Whether it becomes a useful shopping habit or another short-lived experiment will depend on the breadth of the rollout and, more importantly, whether buyers trust the results enough to complete the sale.







