Google tests direct Flipkart purchases through Gemini and AI Mode in India
Google is testing direct Flipkart checkout in Gemini and AI Mode for some Indian shoppers, signaling a push from AI shopping search toward transactions.
Google has begun testing a feature that lets some shoppers in India buy selected Flipkart products through Gemini and Google’s AI Mode, marking a move from AI-assisted product discovery toward completed transactions.
The limited experiment adds a “Buy” button to certain Flipkart listings shown in Gemini and AI Mode. Tapping it opens a Flipkart-branded checkout flow without requiring the shopper to begin with a conventional search journey, according to people familiar with the test and an experience observed by TechCrunch.
The trial currently covers a small group of users and a narrow range of products, including smartphones, electronics, and mobile accessories. Google is expected to expand the experience later in October, ahead of India’s major festive shopping season, one person familiar with the rollout told TechCrunch.
What Google is testing
The test is not yet a universal shopping function across Google’s AI products. Some users continue to see ordinary Flipkart product listings in Gemini and AI Mode, while others receive the additional option to buy. The purchasing feature also does not appear to cover every retailer surfaced by Google.
In the experience reported by TechCrunch, Flipkart listings included the Buy button, while products from competing retailers such as Amazon appeared without the same direct-purchase option. The test therefore appears to be a selective commerce integration rather than a broad marketplace layer inside Google’s AI interfaces.
Google confirmed only that it regularly tests new features designed to help people discover and connect with businesses. A company spokesperson declined to provide details about the Flipkart experiment, including its technical architecture, availability, or planned scope. Flipkart did not immediately respond to TechCrunch’s request for comment.
The current flow also appears distinct from the Google-hosted checkout experience the company has previously demonstrated. Earlier this year, Google introduced the Universal Commerce Protocol, or UCP, as an open standard intended to help AI agents interact with retailers across product discovery, shopping, and checkout. Google said eligible products could eventually be purchased through Gemini and AI Mode using a checkout experience hosted by Google.
The Flipkart test instead presents a Flipkart-branded checkout flow. It is not clear whether UCP powers the experiment, whether the transaction is technically embedded in Google’s interface, or whether Google is directing the user into a partner-controlled purchase process while maintaining the AI shopping context.
Why Flipkart and India matter
The partnership gives Google a high-profile commerce test in one of the world’s largest internet markets. India has more than a billion internet subscribers, according to the evidence cited by TechCrunch, and Flipkart competes aggressively with Amazon for online shoppers.
That competition becomes especially important during India’s festive season, when e-commerce companies run some of their largest promotions. A wider October rollout would give Google a chance to test whether AI-generated product recommendations can convert into purchases during a period when consumers are already actively comparing prices, devices, and offers.
The relationship is also more significant than a standard platform integration. Walmart owns Flipkart, while Google invested approximately $350 million in the company in 2024 as part of a funding round led by the U.S. retailer. Google consequently has both a commercial technology relationship and a minority financial stake in the e-commerce company.
Google said earlier in September that Flipkart was among the merchants working with it on what it calls “agentic shopping” experiences in India. The company did not disclose the details or timeline at that point. The newly reported test provides the first concrete indication of how that partnership may reach consumers.
From recommendations to transactions
For AI product teams, the important change is not simply the presence of a Buy button. It is the attempt to connect conversational or generative product discovery with the operational steps required to complete an order.
AI shopping systems must handle more than product descriptions. They need current inventory, price and promotion data, retailer authentication, payment flows, shipping information, returns, and clear responsibility when an order fails. The Flipkart experiment appears to address only part of that chain, and Google has not disclosed how the systems exchange data or divide support obligations.
The selective nature of the test may help Google manage those risks. Restricting access to certain users, products, and merchants allows the company to observe conversion, checkout reliability, product-data accuracy, and customer-service issues before expanding the feature. It also limits the number of transactions exposed to an unproven workflow.
For Flipkart, appearing inside Gemini and AI Mode could provide another route to shoppers at the moment they are expressing purchase intent. For Google, it offers a way to test whether its AI interfaces can retain commercial value after traditional search results are replaced by conversational answers and generated recommendations.
The experiment also raises questions for competing retailers. If Google gives one marketplace a direct purchase action while showing Amazon and other sellers only as standard listings, merchants may have to decide whether deeper integrations are necessary to remain visible in AI-mediated shopping journeys. That could make access to AI commerce systems an increasingly important distribution issue.
What to watch next
The first signal will be whether Google expands the Buy button beyond the reported product categories and test users later in October. A broader rollout would show that the company considers the initial results sufficient to support higher transaction volume.
Observers should also watch whether Google identifies the technology behind the Flipkart flow. Confirmation that the test uses the Universal Commerce Protocol would provide evidence that UCP is moving from a stated standard into a live merchant integration. A separate implementation would suggest Google is testing multiple models for connecting AI interfaces to retailer checkout.
Other indicators include whether Amazon or additional Indian merchants receive comparable purchase capabilities, whether Google hosts checkout or leaves it with retailers, and how returns, payment disputes, and customer support are handled. Those operational details will matter more to enterprise buyers than the interface itself.
Finally, Google’s eventual disclosures should clarify whether the feature is available only in India or is a template for broader regional commerce partnerships. The company’s comments so far establish an experiment, not a confirmed global product launch.
Creati.ai perspective
Google’s Flipkart test is an important but still narrow step in the competition to make AI systems part of the transaction, not just the research process. The limited rollout and undisclosed technical design make it too early to judge adoption or commercial impact.
For builders and enterprise teams, the practical lesson is to treat AI commerce as an integration problem involving data quality, retailer systems, payments, and accountability. A convincing product recommendation may attract attention, but reliable checkout and post-purchase support will determine whether AI shopping becomes a durable channel.
