Cognizant Claude group: 25,000 staff to run AI inside client firms

Cognizant, the Nasdaq-listed IT services company, said on Tuesday, 10 November 2026 that it has formed a Claude Frontier Capability Group of more than 25,000 professionals whose job is to deliver Anthropic's Claude AI models to enterprise clients. Alongside the new unit, the company introduced a managed services model under which Cognizant will run Claude-powered systems in production for clients, rather than only building them and handing them over. The offer bundles dedicated staff, licensed access to Claude and Cognizant's own tools, including what it calls its AI Delivery Operating System.

For many large companies, the hard part of artificial intelligence is keeping AI systems working inside real operations such as claims, supply chains, factory floors and software teams. Cognizant's move is a bet that clients will pay a services firm to take responsibility for that day-to-day running, not just for a one-time build. Its chief executive, Ravi Kumar S, described the new group as a significant bet on Claude and on what enterprises can achieve with it in their core operations, according to the company's statement.

The structure is also notable. By running two units of similar size, one for Anthropic's Claude and one for OpenAI's products, Cognizant is signalling that it does not expect clients to settle on a single model provider. The new group also builds on an earlier arrangement under which Anthropic said Cognizant would make Claude available to up to 350,000 of its employees.

For India's IT services companies, which compete with Cognizant for the same global contracts, the announcement adds pressure to show similar capacity. On the same day, Wipro announced a partnership with US firm Labelbox to move AI agents that learn from enterprise experience out of pilots and into production, linking Labelbox's reinforcement learning tools with its Wipro Intelligence suite. Both moves point the same way: services companies are trying to sell running AI as an ongoing service, which could replace part of the revenue lost as AI automates traditional coding and maintenance work.

The client results Cognizant quoted, 40% to 60% faster processes at the insurer and 35% less downtime at the drug maker, are the company's own figures. It did not name the clients.

Peepals calculation: The new Claude group equals about 7% of Cognizant's workforce (25,000 / 359,900 = 0.069), and together with the OpenAI group it covers about 14% of all staff (50,000 / 359,900 = 0.139). The Claude group alone is about 2.5 times the company's net hiring over the past year (25,000 / 10,100 = 2.48), which suggests it was formed largely by moving existing employees rather than by new hiring. Our calculation from the reported figures.

Short-term: Enterprise clients of Cognizant now have a named team and a managed service through which to run Claude-based systems, with the company taking on operational responsibility.

Long-term: If managed AI services take hold, IT services revenue could shift from billing for people's hours towards charging for AI systems kept running in production. Jobs inside these companies would shift with it, towards staff who design, supervise and maintain AI agents rather than write and test code by hand.

Who is affected: Cognizant's employees, especially those moved into the new group; large enterprises in insurance, pharmaceuticals, manufacturing and other sectors weighing how to run AI at scale; Indian IT services companies such as Wipro competing for the same work; and engineering graduates whose skills will be judged against AI-delivery needs.

Questions and Answers