How Swiggy is building India’s AI-native commerce infrastructure

For nearly two decades, digital commerce has revolved around a familiar ritual: open an app, browse options, add items to a cart, complete payment, and wait for delivery.

Madhusudhan Rao, Chief Technology Officer at Swiggy, believes that ritual is about to disappear. Instead of navigating apps, consumers will increasingly interact with AI assistants capable of understanding intent, making decisions, and completing entire transactions autonomously. Ordering dinner, restocking groceries, booking a restaurant, or planning an evening out could soon happen through a simple conversation with an AI agent—without ever opening the Swiggy app.

“The bigger shift is that commerce is moving beyond apps into every interface that people already use,” says Rao. “A user can ask an AI, ‘Order dinner for tonight, groceries for tomorrow, and book a table for Friday,’ and the entire flow can happen without opening the Swiggy app. That’s where convenience is headed—ambient, conversational, invisible.”

For Swiggy, this represents far more than another technology upgrade. It signals a strategic transformation from operating one of India’s largest consumer platforms to becoming the underlying commerce infrastructure powering AI-native experiences across multiple digital ecosystems.

Commerce is moving beyond apps

Swiggy’s evolution mirrors a broader shift taking place across the technology industry. The company began by solving a single consumer problem—food delivery. Over time, it expanded into quick commerce, grocery delivery, dining experiences, and hyperlocal services. Yet Rao argues that these are merely different expressions of the same underlying need: enabling everyday consumption.

Consumers do not think in terms of product categories or business verticals. They simply want something accomplished with minimal effort. Artificial intelligence is rapidly changing how those needs are fulfilled.

Instead of navigating multiple applications, users will increasingly express their intent through conversational interfaces embedded inside cars, smart devices, enterprise workflows, messaging platforms, or AI assistants. Commerce, in this future, becomes an invisible layer operating seamlessly in the background.

Rather than asking users to come to Swiggy, the company wants Swiggy to be available wherever customers already are.

Becoming the infrastructure behind AI commerce

This vision explains the strategic significance of Swiggy Builders Club, the company’s recently launched developer initiative.

While it may appear to be another developer ecosystem on the surface, Rao believes its role is far more fundamental. “The real question in the AI era isn’t which app wins,” he says. “It’s which infrastructure layer AI agents choose to transact on.”

Builders Club opens Swiggy’s commerce capabilities to developers and enterprises through the Model Context Protocol (MCP), allowing AI-native applications to integrate food ordering, grocery shopping, and dining reservations directly into intelligent workflows.

The early response has been encouraging. More than 2,000 developers have already applied to join the initiative.

But Rao believes the real value lies beyond developer participation. Instead of spending years building merchant networks, logistics capabilities, and commerce infrastructure from scratch, developers gain immediate access to a mature production ecosystem spanning over 700 cities, more than 270,000 restaurant partners, over 50,000 Instamart products, and millions of monthly users.

In effect, Swiggy is repositioning itself from a consumer-facing application into the transaction layer powering an emerging ecosystem of AI-driven commerce.

From chatbots to autonomous commerce

Much of today’s AI remains conversational. Users ask questions, and systems generate responses. Rao believes the next generation of AI will go much further.

Instead of merely answering queries, intelligent agents will execute complete workflows autonomously.

A simple instruction such as “Reorder my usual from last Tuesday” could trigger an AI agent to identify previous purchases, recommend suitable alternatives if items are unavailable, complete checkout, process payment, and monitor delivery—all without requiring further user intervention.

“What makes this work is composability,” Rao explains. “Every tool is structured to chain cleanly into the next one, so the AI layer behaves less like a chatbot and more like an orchestration engine.”

The distinction is significant. Traditional digital experiences required users to navigate interfaces. Agentic commerce shifts the burden to software, allowing intelligent systems to coordinate multiple services behind the scenes while presenting customers with a single, effortless interaction.

For enterprises, this represents a fundamental redesign of customer engagement rather than simply another AI feature.

AI as a tool for inclusion

One of the most compelling applications of this approach extends beyond convenience: Accessibility.

Many accessibility-focused AI agents already interact with Swiggy through voice and conversational interfaces built on top of its MCP infrastructure.

Instead of navigating complex menus or touchscreens, users can simply say, “Order my usual” or “Get this week’s groceries,” while the AI completes the remaining workflow.

For elderly users, people with visual impairments, or individuals uncomfortable with traditional digital interfaces, this significantly lowers the barriers to digital commerce. “Instead of learning how to use an app, people can simply express what they need,” Rao says.

In a country where digital literacy varies widely, conversational AI has the potential to make technology more inclusive by adapting to users rather than forcing users to adapt to technology.

Building once, innovating repeatedly

Behind Swiggy’s AI ambitions lies a less visible but equally important engineering transformation.

Over the past two years, the company has focused on building reusable technology primitives, shared infrastructure, and common developer platforms capable of supporting multiple businesses simultaneously.

Rather than developing every new service independently, Swiggy increasingly treats technology as a collection of modular capabilities that can be recombined rapidly to launch entirely new consumer experiences.

This architectural shift has enabled products such as Toing and Crew to move from concept to market within weeks using relatively small engineering teams while maintaining the scale and reliability expected from a platform serving millions of customers.

The result is not merely faster software development. It is an organisation capable of experimenting, iterating, and launching new businesses at a pace that would have been difficult under traditional product development models.

The future is AI-native commerce

Over the next few years, Rao’s priorities are clear. Swiggy intends to deepen its MCP capabilities, support a growing number of AI-powered interfaces, expand enterprise integrations, and enable richer autonomous workflows for developers and businesses.

The ambition is straightforward yet profound. Swiggy no longer wants to be known only as the app consumers open when they feel hungry. It wants to become the invisible commerce layer that intelligent systems rely on whenever consumers express an intent to buy.

That vision reflects a broader shift underway across the digital economy.

The future of commerce may no longer belong to the platforms with the most downloads or the most engaging user interfaces. Instead, it may belong to the companies whose infrastructure enables AI agents to transact seamlessly across every digital surface.

In that future, consumers may never think about the technology enabling their purchases. And that, perhaps, is precisely the point.