Vodafone Idea (Vi) has launched Vi Intelligence, an AI programme aimed at two things: how the operator runs internally, and what it sells to enterprise customers. The centrepiece is the Vi Intelligence Lab, built with PwC India, which Vi describes as a place to develop, incubate and scale enterprise AI solutions.
What Vi announced
PwC India will put a dedicated Centre of Excellence team inside Vi's Mumbai office. That's the detail worth noticing. A co-located team means people doing the work day to day, not an advisory deal that goes quiet after the launch.
The Lab has three pillars:
- Agentic AI, aimed at business processes that run with increasing autonomy
- Assistive AI, for tools that support people rather than replace steps
- Sovereign and responsible AI, covering data, IP and how models are handled
Every use case has to clear GRACIE, Vi's own Responsible AI framework, before it goes further.
Vi CTO Jagbir Singh said the company's business has always been about connecting people, businesses and devices, and that "the future belongs to intelligence." Vinish Bawa, who leads PwC India's TMT practice, framed it around enterprises moving to intelligent and autonomous processes and needing trusted ecosystems to get there responsibly.
How the Lab is supposed to work
Vi lists several programmes rather than one product:
- Incubation projects aimed at hard business problems
- A track for building proprietary AI and technology capabilities
- Vi AI Open hackathons and challenges drawing in startups
- A "transformation AI factory" to move ideas from experiment to pilot quickly
- AI skilling programmes for the workforce
Why it matters
Indian telcos have spent years trying to sell more than connectivity. Agentic AI gives them another pitch. They already own the network, data centres and enterprise relationships, so they can offer customers a governed place to test agents before those agents touch production.
Enterprise buyers should look hardest at the governance gate. A lot of agentic pilots stall because nobody has decided who signs off on an agent's actions, what data it can use, or how it gets audited. Running every use case through a named framework at least starts with that question.
Delivery is what's still unknown. Labs are easy to announce. The real test is how many pilots reach production with a paying customer in the next year, and whether GRACIE is published in enough detail for customers to check it against their own risk policies.
Source: Vi bets on agentic AI with new Intelligence Lab built alongside PwC India, Express Computer
