Flipkart is moving AI from a support tool to a core part of its business. The company says roughly 35% to 40% of its internal code is now being produced by AI tools.
That figure signals a major shift in how the e-commerce platform builds products and systems. Instead of waiting to see whether AI will matter, Flipkart is investing deeper and building its own large language model for e-commerce-specific tasks.
Why Flipkart is building its own model
Balaji Thiagarajan, Flipkart’s chief product and technology officer, said the company is working on an “agentic e-commerce platform.” He said the platform combines frontier AI models with Flipkart’s own models.
His comments point to a strategy that goes beyond general-purpose AI. Flipkart believes off-the-shelf models alone are not enough for the highly specific demands of e-commerce.
Thiagarajan argued that the company needs its own model so AI can perform better on specialized tasks. He said that model specialization is becoming one of the key differentiators in AI deployment.
| AI Usage Area | How Flipkart Uses It | Notes |
|---|---|---|
| Internal code generation | 35% to 40% of code | Produced by AI tools |
| Product discovery | AI-supported shopping experience | Part of customer-facing use cases |
| Conversational shopping | AI-assisted interactions | Built for e-commerce workflows |
| Seller operations | Operational support | Used across the seller ecosystem |
Flipkart says it has deployed more than 250 models across its ecosystem. Those models are used in areas ranging from product search to internal productivity.
AI is reaching more than coding
The company’s AI push extends beyond software development. It is being used for product discovery, conversational shopping, seller operations, and internal work across the company.
One of the most striking applications is code writing. Thiagarajan said around 35% to 40% of the company’s code is now generated by AI tools.
The trend reflects a broader shift in the technology industry, where more engineering work is being handed to AI systems to speed up development.
In a similar case, Meesho has reportedly generated more than 70% of its code with AI. OpenAI president Greg Brockman has also said AI wrote about 80% of the code at the startup.
A mixture of experts, not one dominant model
Thiagarajan said the future will likely involve different AI models for different jobs. He suggested Flipkart may combine external tools such as Claude Code with its own models.
He described that direction as a “mixture of experts.” In that setup, some experts come from outside models while others are built in-house by Flipkart.
According to him, the real differentiation will come from data, engineering strength, and the ability to build a highly specific e-commerce LLM. The value lies not only in access to AI, but in how that AI is trained and applied to business problems.
That approach explains why Flipkart is investing in narrower, more targeted models. For tasks such as product search, shopping conversations, and seller workflows, the company sees domain accuracy and context as more important than broad general capability.
Still investing, with governance in focus
Despite rising AI costs across the industry, Flipkart says it is not focusing heavily on ROI at this stage. Thiagarajan said the company remains in investment mode.
He added that the bigger current focus is AI governance. The main areas include content moderation, response fidelity, human-in-the-loop systems, and reinforcement learning.
That emphasis matters because AI is being used in functions that interact directly with customers and sellers. Incorrect answers, unsuitable content, or unchecked automated decisions become more serious as AI usage grows at scale.
Flipkart’s organizational changes have followed the same direction. The company recently appointed Vinay Vaidya as SVP, Technology for Supply Chain.
The move comes as AI investment competition in India intensifies. Amazon, one of Flipkart’s main rivals, has announced a $13 billion investment in India through 2030 to expand cloud and AI infrastructure, bringing its total planned investment in the country to $48 billion.
While companies such as Uber and Walmart are reducing AI use for employees because of rising costs, Flipkart is continuing to raise its bet. Its focus is not simply on cutting costs, but on building an AI foundation strong enough to support e-commerce operations at a larger scale.