Flipkart is leveraging AI to move beyond keyword-led product searches to interpret what shoppers want, ask follow-up questions, and adapt recommendations to their behaviour while they browse.

Speaking at Inc42’s inaugural ‘The CTO Summit 2026’, Flipkart CPTO Balaji Thiagarajan said the company’s approach to personalisation brings together search, conversations, voice, visual inputs, and content feeds.

“Search was going from keyword search to four words and six words, which was slowly starting to become sentences. So our search stack became AI-native,” he said.

Thiagarajan described Flipkart AI Mode as an evolution of the search box that can move between transactional queries and conversations. When a shopper’s request is broad, the system can ask follow-up questions to narrow down their requirements. Shoppers can also use voice or visual inputs to explain what they want.

The company is also adapting browsing feeds to interactions within a shopping session. If a shopper engages with shoes, for instance, subsequent recommendations can show more footwear.

Thiagarajan said the aim is “N=1” personalisation — an experience tailored to each individual. Building this requires understanding customers’ interests and continually adapting the platform experience, while handling their data responsibly, he added.

Building For Different Languages 

Serving customers across India, including more than 22,000 pincodes, also means accommodating different languages and ways of interacting with the platform, Thiagarajan said.

He pointed to customers in Tier II and III cities who manage household purchases but may not be comfortable shopping in English. Flipkart’s response has been to support multiple languages and introduce alternatives to typing.

“Then we asked, ‘is typing on a phone the only way to do this?’ So we went multimodal. Now voice is a form of interaction. You can take a picture or look at a video and say, ‘This is something I want to buy,’ and the system should be able to understand,” he said.

From Finding Products To Delegating Purchases

Thiagarajan said Flipkart’s long-term goal is to become the customer’s “personal concierge for shopping”, available through their preferred channels and at the times they choose.

He outlined potential agent-led use cases, including finding products based on a shopper’s requirements, scouting for deals that match a preferred brand and price, and replenishing frequently purchased household staples.

Consumers could choose how much to delegate based on their comfort, ranging from assistance with shopping to autonomous actions, he said. However, “control will always be with the consumer”.

Earlier in the day, Swiggy CTO Madhusudhan Rao also outlined that the company is expanding access to its commerce services through external AI agents, enabling users to discover food, build carts, place orders, and book restaurant tables outside its own app.

The post How Flipkart Is Using AI To Move Beyond Keyword-Led Product Discovery appeared first on Inc42 Media.