Erik Ekkuden, chief technology officer at the Swedish network equipment maker Ericsson, told Business Standard that he expects India to lead the world in adopting agentic AI — and in democratising it. The claim is worth taking seriously less as prophecy than as a signal of where a major infrastructure vendor sees demand forming. Agentic AI, in the current shorthand, refers to systems that do not merely answer questions but carry out tasks: planning a sequence of steps, calling other software, correcting course when something fails. The difference between a chatbot and an agent is roughly the difference between advice and delegation.

India is a plausible candidate for the role Ekkuden assigns it. The country has a record of skipping technology generations — large parts of the population moved from limited landline access straight to mobile, and from cash to UPI payments at a pace that surprised outside observers. Its IT services industry, built on running other people's systems at scale, understands enterprise integration from the inside. And democratisation, in this context, is not a vague slogan. An agent that can work in Indian languages, over voice, on an inexpensive phone, could serve a small merchant managing suppliers and payments — a user that English-first, text-heavy tools never properly reached.

The supply chain catch

Ekkuden's caveat is the more interesting half of the statement: the challenge, he said, will lie in building the supply chain ecosystem. Agentic AI at national scale is not just a software story. It rests on compute, data centres, chips, devices and the networks connecting them — and India remains heavily import-dependent on much of that hardware. Its ambitions in domestic semiconductor manufacturing are real but early, and a lead in adoption can easily outrun the industrial base underneath it. That pattern has a history: the markets that adopt fastest are not always the ones that capture the value of what they adopt.

It is also worth noting who is making the prediction. Ericsson sells network equipment, and autonomous agents are, from a network operator's perspective, a traffic story: machine-to-machine calls, continuous API activity, interactions that need to be fast and reliable. If agents proliferate, someone bills for the connectivity, and Ericsson intends to be that someone. Optimism from a vendor whose equipment the vision depends on should be read with that in mind — though a commercial stake does not make a forecast wrong.

What leading would actually require

Adoption leadership is harder than it sounds. An agent that misbooks a ticket is an annoyance; an agent that mispays an invoice or commits a company to a contract raises questions of liability and trust that the chatbot era never had to answer. That means redress mechanisms, audit trails, and the unglamorous work of integrating with legacy systems — areas where India's services industry is arguably well placed, but where no country has yet set the standard. Scale cuts both ways: what works in India will work at enormous volume, and so will what fails.

The prediction is testable, which is more than can be said for most technology forecasts. Watch whether agents show up in vernacular, small-business settings rather than staying confined to enterprise pilots, and whether supply chain talk turns into fabs, packaging plants and locally available compute. If both happen over the next few years, Ekkuden will have called something real. If only the first does, India will have demonstrated something slightly different: that it can adopt the world's software faster than it can build the machinery underneath it — its own kind of leadership, with its own kind of dependence attached.