Agentic commerce must be transparent and accountable, says InMobi CEO

InMobi Group CEO Naveen Tewari says agentic commerce can reshape buying behaviour, supply chains and commerce journeys, but warns these AI-driven systems must stay transparent and accountable

InMobi Group CEO and founder Naveen Tewari said agentic commerce holds great potential to change how people shop but the system needs to be transparent and accountable.

 

"If you bring AI intelligence into commerce, you fundamentally shift economic growth. But this must be authentic. Digital platforms over the past decade have created distortion. Social media has often manipulated attention. Agentic systems must be transparent and accountable," he said at the AI Impact Summit.

 

He added that when intelligence becomes embedded in commerce, it impacts people's buying behaviour, supply chains and manufacturers. That will help companies to move to personal feeds from personalised feeds.

 

“What you would see here is personal feeds of consumers getting created in real time with products on them,” he said. Tewari added that the company plans to train commerce models at massive scale over the coming years, with the objective of bringing intelligence into the commerce journey for large numbers of consumers.

 

Explaining the architecture behind Glance’s approach, Tewari said multiple models come together to create the agentic experience. He described a “commerce intelligence graph” akin to a knowledge graph. It captures commerce elements and understanding, and a generative AI experience layer that produces output designed for commerce use cases.

 

Glance, which houses the consumer businesses of InMobi, counts Google, Jio Platforms and Mithril Capital among its investors besides parent InMobi, which holds about 50 per cent. It has so far raised $390 million at a $1.7 billion valuation. InMobi Group has already invested $200 million in Glance AI and a generative AI ad tech stack.

 

A key element in the agentic commerce era will be a “living commerce context graph” that understands a person’s context in the moment, what they are seeking, and their price and brand sensitivity.

 

“You can actually put that into the model, and it’ll find out the pathway for the most efficient buying for you,” he said, describing it as purchase-path optimisation that can evaluate “millions and billions” of possible pathways.