TLDR;
Google is reportedly testing a Buy button on selected Flipkart listings in Gemini and AI Mode for a small group of users in India. The button opens a Flipkart-branded checkout inside Google’s AI interface, but Google has not announced the test, its underlying technology or a firm rollout. Brands should treat this as an early signal that access to AI checkout may depend on merchant integrations, clean product data and clear attribution.
What happened
TechCrunch reported the test on 26 September after seeing the experience and speaking with people familiar with it. Some users in India can see a Buy button on selected Flipkart listings for smartphones, electronics and mobile accessories inside Gemini and AI Mode. Tapping it opens a Flipkart-branded checkout without leaving the AI interface. The report said access is limited and could expand in October, but Google only confirmed that it regularly tests ways to help people discover and connect with businesses. Flipkart did not comment.
The checkout flow matters because it appears different from the Google-hosted checkout previously shown for the Universal Commerce Protocol. The observed flow carries Flipkart branding, and neither company has said which technology powers it. Google had already named Flipkart as an agentic-shopping partner in India, but that does not prove the test uses UCP or that other participating merchants will receive the same Buy button.
Google had already added product listings, comparison tables, prices and places to buy in Gemini and conversational shopping results in AI Mode in India. This test adds a transaction step to that discovery journey. Listings from Amazon reportedly appeared in the same experience without a Buy button, which suggests checkout eligibility is selective rather than a standard feature for every retailer surfaced by Google.
Why it matters
The commercial shift is simple: Google can now sit across discovery, comparison and the handoff into checkout. Flipkart keeps its branding and appears to operate the transaction, while Google controls the conversational surface where the product is selected. For retailers, that creates a new conversion path with fewer page visits before purchase and less room to recover from weak product data or an uncompetitive offer.
It also complicates measurement. A customer may compare several merchants in an AI answer, tap Buy and complete a partner-branded checkout without a conventional landing-page session. Enterprise analytics teams need to know which identifiers survive the handoff, who records consent, how assisted conversions are attributed and whether the transaction is visible in existing channel reporting. Without that detail, a promising checkout volume can still distort acquisition and incrementality reporting.
The test should not be treated as a new ranking factor or an open merchant programme. Google has a financial relationship with Flipkart and the trial is narrow. Selection could reflect a commercial integration, product eligibility, technical readiness or test design. Until Google publishes rules, brands should avoid assuming that UCP support or Merchant Center participation alone will unlock the same experience.
How your brand can benefit / be affected
Start with product truth. Check that identifiers, variants, prices, availability, delivery promises and returns information match across Merchant Center, your product APIs and checkout. AI shopping makes contradictions more visible because the assistant may compare a product card with live checkout facts in the same conversation. A stale price or unavailable colour can break trust before the customer reaches your site.
Prepare measurement for a checkout that begins outside the normal storefront journey. Ask platform and commerce teams to preserve a source identifier through cart creation, payment and order records. Define how server-side conversion events, analytics sessions and affiliate or partner credits will reconcile. Measure completed orders, cancellations, returns and margin, rather than treating Buy-button clicks as proof of demand.
Set commercial and service ownership before joining any similar pilot. Confirm who is merchant of record, who handles payment failure, fraud, tax, delivery changes, returns and customer support, and what product-ranking or offer rules apply. Require a clear path for correcting catalogue errors and pausing checkout when stock, price or policy data is unreliable.
Monitor the India test, but do not engineer around an undocumented interface. Keep feeds and structured product data accurate, make checkout fast on the retailer’s own properties and test AI referrals separately from ordinary search. If Google opens eligibility, brands with reliable catalogues, transaction APIs and clean attribution will be able to evaluate it quickly without betting the wider commerce stack on a limited experiment.
News date: 26 September 2026. Editorial review: 29 September 2026. Analysis includes subsequent developments where stated.