Google tests buying from Walmart-owned Flipkart through Gemini and AI Mode in India

You have probably watched a colleague ask ChatGPT for a product recommendation, then opened a new tab to buy it on Amazon. The model suggests. The human switches context. The purchase happens somewhere else. Google is now testing whether that handoff can disappear entirely by letting users in India shop Walmart-owned Flipkart without leaving Gemini or the company's AI Mode search interface.

The experiment matters because it collapses two behaviours that have lived apart since the first search engine: finding an answer and completing a transaction. For decades, search sent you to a retailer. Conversational AI was supposed to keep you in dialogue, not send you to a checkout. This pilot fuses both. A user asks Gemini for a laptop or a pair of headphones, the model surfaces Flipkart inventory and the purchase can close inside the same session. No tab-switching. No cart abandonment because you got distracted by email. The friction that e-commerce teams spend millions to reduce – the gap between intent and conversion – shrinks to near zero.

What this means for L&D and digital transformation

Learning and development leaders who have spent the past two years teaching prompt engineering and responsible AI use now face a second-order question: what happens when the tools employees learn to interrogate also become the tools they use to act? A conversational interface that recommends, compares and completes a purchase is no longer a research assistant. It is a decision agent with spending authority. That changes the capability mix your people need.

Training programmes built around AI literacy have rightly focused on how to evaluate model output, spot hallucination and write effective prompts. Those skills remain foundational. But when the model can execute a transaction, employees also need to understand commercial intent, data provenance and the economic relationships that shape what the AI recommends. If Gemini surfaces Flipkart results because Google and Walmart have a commercial arrangement, that is not a neutral search result. It is a curated marketplace. Your teams need to recognise the difference and decide when to trust the curation and when to step outside it.

The shift also has implications for procurement, finance and compliance functions. If a conversational AI can buy on behalf of a user, organisations will need guardrails that govern spending limits, approval workflows and vendor eligibility inside the chat interface itself. The controls that live in your ERP or procurement platform today will need to extend into the conversational layer. That is not a technical problem alone. It is a process design and change management challenge. L&D will be asked to help employees navigate systems where the boundary between asking a question and committing budget has become porous.

The broader pattern

Google's test in India is part of a wider pattern. OpenAI has discussed plugins that let ChatGPT book travel or order food. Amazon has built Alexa commerce for years, though adoption has been modest. Microsoft is weaving Copilot into enterprise workflows where a prompt can draft a document, schedule a meeting or submit an expense claim. The common thread is that conversational AI is moving from advisory to transactional. The model does not just inform your decision. It executes it.

For innovation and transformation leaders, the question is not whether this shift will reach your organisation. It is how quickly and in what form. The capabilities your people need to work effectively with AI are no longer static. They are evolving as the tools themselves gain agency. A learning strategy that treats AI as a research layer will miss the moment when it becomes an execution layer. The organisations that adapt fastest will be those that see this transition coming and start building the skills, governance and process changes it demands before the tools arrive at scale.

The Google-Flipkart pilot is a test in one market with one retailer. But the principle it demonstrates – that conversational AI can collapse the gap between question and transaction – will not stay confined to consumer shopping in India. It will spread to enterprise procurement, internal service requests and any workflow where a human currently translates an AI recommendation into action. The handoff is disappearing. The capabilities your people need must keep pace.

Sources:

• TechCrunch, 26 September 2026