Google is quietly shifting its AI strategy in India from product discovery to direct transactional execution. Select users of the Gemini app and Google’s AI Mode in Search are now seeing a “Buy” button on Flipkart product listings, enabling a direct checkout flow entirely within the AI interface.
This pilot, first reported by TechCrunch, represents the first major consumer-facing implementation of Google’s Universal Commerce Protocol (UCP). The timing is deliberate. Google is aiming for a wider rollout in October, just ahead of India’s high-stakes festive shopping season, when e-commerce giants Flipkart and Amazon engage in their most aggressive annual sales campaigns.
The Mechanics of the In-App Checkout
For the selected cohort of testers, the shopping experience bypasses the traditional redirect loop. When a user asks Gemini for product recommendations—such as “What are the best smartphones under Rs 20,000?”—the AI generates a list of options. For eligible Flipkart listings, a branded “Buy” button appears alongside the product details.
Tapping this button does not redirect the user to the Flipkart app or web storefront. Instead, it launches a Flipkart-branded checkout overlay directly within the Gemini interface. The transaction, from address confirmation to payment processing, is completed without leaving the conversational flow.
Currently, the pilot is restricted to a narrow catalog of high-demand categories:
- Smartphones
- Consumer electronics
- Mobile accessories
This implementation differs from the Google-hosted checkout experience the search giant demonstrated earlier this year. Rather than keeping the entire transaction within Google’s proprietary payment rails, this “embedded option” allows Flipkart to maintain its own branding, customer account integration, and payment processing flow.
The Universal Commerce Protocol (UCP) Engine
At the core of this integration is the Universal Commerce Protocol (UCP), an open-source standard designed to power agentic commerce. Introduced in early 2026, UCP acts as an open, standardized translation layer between a merchant’s backend inventory and autonomous AI agents.
According to Google’s technical documentation on GitHub, UCP establishes a common language for inventory discovery, cart management, and secure checkout. It is designed to work with existing retail infrastructure and is compatible with the Agent Payments Protocol (AP2) to ensure secure, cryptographically verified transactions.
| Traditional E-Commerce | Agentic Commerce (UCP) |
|---|---|
| Destination-based (Websites/Apps) | Utility-based (AI interfaces) |
| Centralized walled gardens | Decentralized, protocol-driven |
| Manual multi-step checkout | Autonomous, in-context transaction |
| High cart abandonment rates | Instant, frictionless conversion |
Rather than requiring custom, one-off integrations for every AI platform, retailers who adopt UCP can declare their capabilities dynamically. This allows any compatible AI agent to query real-time stock levels, apply personalized loyalty discounts, and initiate checkout sessions securely.
The Amazon Gap and Strategic Alliances
The pilot highlights a growing divide in how major platforms approach the agentic era. While Flipkart listings inside Gemini feature the direct checkout option, competing listings from Amazon do not. Amazon products still appear in Gemini’s comparison tables, but they remain static “blue links” that require the user to exit the AI interface to complete a purchase.
This asymmetry is partly driven by corporate alliances. Google holds a direct financial interest in Flipkart, having invested approximately $350 million in the Walmart-owned e-commerce leader during a 2024 funding round. This financial and technological partnership has positioned Flipkart as the primary launch partner for Google’s agentic commerce experiments in India, the world’s second-largest internet market with over one billion subscribers.
For Amazon, the decision to remain outside the direct checkout loop reflects a broader strategic tension. Allowing third-party AI agents to handle transactions threatens to turn Amazon’s highly optimized storefront into a background utility, stripping the company of direct customer touchpoints, advertising impressions, and cross-selling opportunities.
Disruption of the Affiliate Middleman
The transition from conversational search to direct transaction introduces immediate friction for the broader digital ecosystem, particularly affiliate marketing and product review platforms.
In the traditional web economy, publishers, tech bloggers, and coupon sites act as the bridge between discovery and purchase. A consumer reads a review, clicks an affiliate link, and the publisher earns a commission on the resulting sale.
If Gemini can recommend a product, compare specifications, and execute the purchase natively, the user has no incentive to visit external review sites. This bypasses the traditional affiliate link loop entirely. If the transaction occurs within an AI chat, it remains unclear how, or if, content creators who originally trained the models with their reviews will be compensated.
The Security and Trust Hurdle
Beyond the economic implications, agentic commerce faces significant technical and security challenges. Executing payments via AI agents requires robust authentication to prevent unauthorized purchases.
Google’s framework relies on Device Bound Session Credentials (DBSC), a W3C standard that cryptographically links active login tokens to a user’s physical hardware. This ensures that even if an AI session is intercepted, the transaction cannot be hijacked by unauthorized actors.
As Google prepares to expand the Flipkart pilot in October, the ultimate test will be consumer trust. Moving from “What should I buy?” to “Buy it for me” requires users to hand over transactional control to an AI assistant. For now, Google’s hybrid approach—using a Flipkart-branded checkout inside Gemini—serves as a transitional step, balancing the friction-free promise of agentic commerce with the familiarity of established retail brands.
