Google Flipkart checkout is being tested inside Gemini and AI Mode for a limited group of shoppers in India, moving Google’s conversational shopping tools from product discovery toward transactions. Eligible listings show a Buy button that opens a Flipkart-branded checkout flow without sending the shopper to a separate browsing experience.

 

The experiment was first reported after TechCrunch observed the flow and spoke with people familiar with the test. Google confirmed that it routinely tests ways to connect users with businesses, but did not disclose the technology, participating users or a public launch schedule.

 

The reported test has three important limits:

  • Only selected Indian users can see the Buy button.
  • The initial catalogue covers smartphones, electronics and mobile accessories.
  • A broader October rollout is reported, not formally announced.

 

Google Flipkart Checkout Test Is Deliberately Narrow

The distinction between a test and a launch matters. Most Gemini and AI Mode users continue to see ordinary Flipkart product listings, while only participants in the experiment receive a direct purchase option. That makes the feature an early commercial trial rather than a generally available service.

 

For participating users, the new step appears at the point where a recommendation becomes a purchase. The Buy button opens a checkout page carrying Flipkart’s branding inside Google’s AI interface, according to the experience reviewed by TechCrunch and summarized by The Indian Express.

 

The report says Google is preparing wider availability later in October, ahead of India’s festive shopping season. Neither Google nor Flipkart has published eligibility rules, supported payment methods, return handling, seller requirements or the exact date of any expansion.

 

Those omissions prevent firm conclusions about scale. They also leave open whether the checkout relies on Google’s Universal Commerce Protocol, a direct Flipkart integration or another technical arrangement. Google’s spokesperson offered no additional detail when asked about the trial.

 

Gemini and AI Mode Move From Answers to Transactions

Google has already described native checkout as an early use for its Universal Commerce Protocol, or UCP. Announced in January, the open standard is designed to let AI systems communicate with merchant websites and payment services across the journey from product discovery to post-purchase support.

 

Google’s initial native-checkout announcement focused on eligible US retailers and Google Pay. The Flipkart test is different because the visible checkout is Flipkart-branded and is being tried in India. That suggests Google is testing how its broader commerce strategy can adapt to a major local marketplace rather than simply copying the US implementation.

 

The commercial logic is straightforward. Search and chatbot interfaces can influence which products shoppers consider, but a completed checkout gives Google and participating merchants a clearer path to conversion. It also reduces the number of taps between an AI-generated recommendation and a sale.

 

That convenience increases the importance of ranking rules and disclosure. A shopper needs to know whether a product appears because it is the best match, a paid placement, an exclusive integration or the only offer capable of completing checkout inside the interface.

 

Background Reading

 

India Gives Google a High-Volume Commerce Test Bed

India combines a large mobile-shopping market with unusually broad adoption of real-time digital payments. The country’s Unified Payments Interface processed 24.51 billion transactions worth 29.82 trillion rupees in August, according to Reuters.

 

India’s payments authority is separately preparing infrastructure for agentic payments, including spending limits, audit trails and identity checks. The Flipkart test should not be assumed to use that system: Google and Flipkart have not disclosed the payment rail or the level of automation behind the checkout.

 

The limited product mix also appears purposeful. Smartphones and accessories are searchable, specification-heavy categories where shoppers already compare prices and features online. They offer Google a controlled way to study whether conversational recommendations can lead to purchases without immediately covering groceries, fashion or other categories with different fulfillment and return patterns.

 

Flipkart gives the experiment local reach and merchant inventory, while Gemini and AI Mode supply the discovery interface. If the October expansion occurs, the practical test will be whether shoppers trust the integrated route enough to complete payment rather than opening a marketplace app to compare listings independently.

 

Checkout Raises Bias, Fraud and Accountability Questions

The checkout trial arrives as banks are warning that AI shopping agents are advancing faster than consumer protections. A group including Bank of America, NatWest and ING said agents could expose card data, steer users toward weaker payment protections or leave customers uncertain about who is responsible when a purchase goes wrong.

 

Independent research also shows why recommendation transparency matters. A recent audit of 1,536 chatbot and search responses found that displayed sources varied sharply across interfaces and repeated queries. The researchers argued that consumer-facing AI shopping should be evaluated under the actual conditions users encounter, not inferred from isolated API tests.

 

Google’s narrow rollout gives it room to test these failure points before broad availability. Useful disclosures would include why a listing was selected, whether placement is sponsored, which company holds payment credentials, how consent is recorded and who handles refunds, disputes and unauthorized purchases.

 

The reported October expansion remains the next verifiable milestone. Until Google or Flipkart publishes terms and availability, the strongest conclusion is limited but significant: Google is testing a transaction layer inside Gemini and AI Mode, with Flipkart checkout serving as its first visible Indian implementation.