Naveen Tewari on AI Shopping Agents and the Future of Commerce

In a July 2026 sponsored episode of The AI Download, Shira Lazar interviews Naveen Tewari, founder and CEO of InMobi and Glance. The companies sponsor the episode. They discuss his business history and vision for AI shopping agents.

Tewari credits persistence and luck for entrepreneurial success. He says the company evolved through repeated experiments and that its present form was not predictable when it began.

He recalls building InMobi in India in 2007 and 2008, when he says working at a startup was less attractive to many engineers. He describes the difficulty of recruiting and the skepticism the founders encountered.

Tewari says India’s startup ecosystem later grew far beyond what he expected. He describes entrepreneurship as a new source of aspiration for younger people and takes pride in his company’s early role.

Asked about becoming a unicorn, he explains that the term refers to a startup valuation above $1 billion. He recalls not initially recognizing the term when people congratulated him.

He describes InMobi as an advertising platform that helps developers monetize mobile applications. In his account, advertising made many digital services accessible without a direct user fee.

Tewari argues that advertising will similarly fund broad access to AI services. Shira raises the consumer tradeoff between paying directly and receiving ads; his forecast reflects the interests of an advertising executive.

He describes Glance as a consumer venture that followed several experiments. His aim, he says, is to use AI for shopping rather than build a frontier model.

Shira says she tried the app and found that it generated outfit ideas. She compares the experience with asking a general AI assistant for styling help.

Tewari divides shopping into discovery and selection on one side, and purchase on the other. He describes a system that learns preferences and presents personalized visual suggestions.

His example combines an image of a shopper with a sweater so the person can decide whether they like the look. He presents this as an easier way to react to a recommendation than reading text alone.

He then describes an agent finding purchase options based on price and preferences. These are his description and aspirations for the service, not independent evidence that it always finds the best option.

Shira asks how this differs from general AI shopping. Tewari says prompting can be difficult and presents a visual-first interface as his company’s answer.

Shira describes uploading a photo and seeing professional, artistic and location-themed looks. She says some matched her taste and others gave her new ideas, with options to share or find products.

Tewari compares the suggestions to a trusted friend encouraging someone to try a different style. Shira calls it a combination of a personal shopper and stylist.

He says fashion was a starting category because it produces detailed preference signals, and describes plans for accessories, beauty, pets and travel. These are company plans at the time of the interview.

Tewari says commerce is difficult to change because physical goods and supply chains are involved. He predicts that agents could change both product discovery and purchasing.

He claims better selection could reduce returns and expose shoppers to smaller brands. He also imagines agents combining price, discounts and loyalty information; the interview does not demonstrate those outcomes at scale.

Shira asks whether smarter purchasing could reduce waste instead of simply encouraging more consumption. Tewari agrees and predicts benefits for supply chains and small producers, while saying he does not have a figure for current inefficiency.

Shira asks how a founder balances organic recommendations with paid placement. Tewari says AI services have inference costs and need a business model, including compensation connected to recommendations or transactions.

He proposes disclosing commissions and explaining when economic incentives affect recommendations. Shira asks whether transparency can build trust; a disclosure alone is not tested in the episode as proof of unbiased results.

The pair discuss data use. Tewari claims information provided through his service is used in a privacy-compliant way to serve the user. The interview does not independently audit that claim or detail all data flows.

Shira asks about shopping on devices beyond phones. Tewari describes a possible television experience in which a viewer selects an item from a scene, sees it visualized and chooses whether to buy.

He envisions the system adapting suggestions as it learns a user’s preferences. He predicts that shopping agents will become common on televisions; this is a forecast rather than an established outcome.

Asked about agent-to-agent advertising, Tewari imagines a recommendation agent and an advertising agent exchanging information. He says the combined system should still act for the consumer.

Shira questions whether an agent could appear helpful while acting against the user’s interests. Tewari returns to transparency about incentives and model behavior as his proposed response.

Tewari says he expects people to approve purchases rather than agents buying without notice. He imagines a future wearable assistant directing a shopper to an item in a nearby physical store.

The speakers discuss the tension between personalization and privacy. Tewari says external scrutiny can keep companies accountable and acknowledges that newer models can infer more from data than earlier systems.

Shira asks whether AI will flatten taste. Tewari argues instead that personalization could help people discover choices, including independent brands; that is his prediction, not a demonstrated universal effect.

Asked what might lose out, he predicts that online marketplaces could face pressure. He argues agents could take over parts of selection and price comparison, while third-party logistics handle fulfillment.

The conversation turns to creators. Tewari imagines a creator training an agent on their taste and receiving a share when their audience uses it for a transaction. Shira connects the idea to concerns about compensation for creative work.

He says his company hopes to launch a creator-focused platform in the future. The proposed revenue model is a plan discussed in 2026, not evidence that creators already receive those payments.

Shira returns to privacy and asks what Glance does differently. Tewari says more data can improve personalization and claims the company intends to use data for the user rather than share it elsewhere.

Asked who should provide guardrails, he says companies bear the greatest responsibility, with government and consumers providing scrutiny. Shira raises younger users’ distrust of corporations.

Tewari says users should be able to make informed choices and argues that companies need to explain their constraints and how they make money. His comments describe a desired approach, not an independent privacy assessment.

Shira asks how people can preserve strategic thinking. Tewari argues that AI can support brainstorming and help people bring ideas to life, while people still choose how to use the tools.

Asked how he challenges himself, he says he surrounds himself with capable people and AI tools. He rejects the idea that using a tool necessarily makes a person lazy.

He repeats two predictions: agents will connect online shopping with physical stores, and the effects may extend to supply chains and manufacturing over several years. Neither outcome is established by the interview.

Tewari imagines a traveler or shopper receiving nearby product suggestions through a wearable device. He identifies travel as another category his company is exploring.

Shira asks about jobs and company culture. Tewari says his employees are expected to use AI and claims the company is producing more; the discussion does not independently verify the productivity figures.

He says faster tools can support a culture of experimentation and urges people affected by job changes to learn to use AI. Shira thanks him and discloses the episode’s sponsorship by InMobi and Glance.