Naveen Tewari on Agentic Commerce, the Future of AI Shopping and Why the Purchase Journey Is About to Change

Naveen Tewari is the founder and CEO of InMobi Group, the global advertising and technology platform behind InMobi Advertising and Glance, an agentic commerce platform scaling rapidly in the United States to 10 million users and counting. Here, he shares what AI actually changes in retail, why shopping was never a text problem and where the purchase journey goes from here.

You’ve been talking about agentic commerce for a while now. What’s your core thesis?

The way I think about it: offline commerce moved to e-commerce. Now, a large portion of the world’s digital commerce will move into agentic commerce. The agent does the hard work of discovery, selection and price optimization on behalf of the consumer. And once that happens, the value proposition of the traditional shopping experience starts to change dramatically. What we’re building is intelligence that sits on top of that journey and makes every step of it more efficient.

Why did you take a visual-first approach to AI shopping? Most of the major AI platforms that are popular today are text-driven.

That’s exactly the problem. If you look at the use cases that the first wave of frontier models are supporting, they’re almost entirely knowledge worker tasks like coding, writing and research. Shopping, on the other hand, is for everyone. And most people struggle to describe what they want in words. How do you explain a dress in text? You can describe some basics like length or neckline but most people are also looking for a vibe or a feeling. The cognitive load alone is enough to make people give up.

So we made a very different bet. Every output Glance generates is visual. And we go one step further: we don’t just show you a product, we show you the product on you, in your environment. For fashion, that means rendered on your own image. For home goods, in your actual home. Our model is trained to understand your features, your coloring and your style signals so that it produces an output that’s genuinely personal to you. That’s a fundamentally different experience from browsing a grid of product images and trying to imagine whether something works for you.

What makes this possible is how differently our models are trained compared to general-purpose frontier models. We’re not trying to do everything — we’re fine-tuned by category. Fashion first, because it’s the most complex: the input consumers give us there is richer and more expressive than almost any other category. Then accessories, beauty, home. Each category has its own model logic, its own way of reading the consumer and generating the right output.

You’re integrating Glance into the television experience. Tell us about that.

Television is where shopping inspiration actually begins for most people — it always has been. The difference now is that we can close the gap between inspiration and action in real time. We’ve partnered with Samsung and Verizon, and what we’ve built is an agent that’s integrated directly into the content experience. A consumer can pause what they’re watching, tap a button, and immediately see a shoppable feed based on what’s on screen — whether that’s a look from a show, a product they noticed in the background or a style moment that caught their attention.

The same product catalog infrastructure that powers the Glance app on your phone also powers the TV experience: 400-plus brands, around 15 million SKUs, with real-time inventory and pricing refresh. So when a consumer acts on that inspiration, they’re seeing what’s actually available and at what price, right now. That connection between the inspiration moment and the live catalog is something that took us about nine months to build properly, and it’s foundational to everything we do across every surface and every device.

What are you seeing in terms of how people actually interact with Glance?

A few things have genuinely surprised us. One is how people engage with the discovery experience versus chat. Only about 20% of users initiate through conversation. The other 80% start visually, moving through the generated feed and responding to what they see. Chat comes in downstream, as more of a refinement layer. Someone sees something close to what they want and then uses the agent to fine-tune it, asking for a different color or pairing, or to adjust the price point. So chat is layer two, not layer one. That’s a meaningful insight about how shopping actually works: it’s a discovery process before it’s an intent-driven one.

The transaction data has also exceeded our expectations. Fifteen percent of users are transacting, and about 40% of daily active users add to cart on any given day. We’re at around 10 million users now and expect to reach 15 million by year-end. At that scale, the engagement signals are telling us something real: when you remove the cognitive friction from discovery and show people products that are genuinely right for them, the path to purchase shortens considerably.

Glance routes users to the brand’s own site to complete the purchase. Why did you design the experience that way?

That was a very deliberate choice. We want to create a relationship between the consumer and the brand, not take that relationship over. What we can do is surface the brand that made exactly the right product for a specific consumer, a brand that person might never have encountered otherwise. A niche label, a smaller label, a brand that doesn’t have the marketing budget to compete for attention on a traditional platform. Our job is intelligence. The brand’s job is everything that comes after. The brands that lean into this early, that let their products be part of how these agentic models get trained, are going to have a meaningful discovery advantage as the category scales.

Enjoyed this article? Eye on the Founder is just one of the ways we bring you closer to the ideas, people, and decisions moving our industry forward. Join our community of senior executives and industry insiders — subscribe to our weekly newsletter here.

About the Author:

Jessica Binns is Editorial Director at Berns & Co., where she leads editorial strategy and content programming, including the launch of a new thought leadership series profiling senior executives shaping the future of fashion, retail, and consumer goods. A journalist and editor with more than 15 years of experience covering apparel, footwear, retail, trade policy and tech, she is a contributing writer for Vogue Business and the former Managing Editor of Sourcing Journal. Her work has also appeared in WWD, Footwear News, and Retail Dive, and she has appeared on CNN This Morning with Audie Cornish to discuss the evolving fashion landscape.