Analysis | Enterprise AI
Cognizant is betting that the most persuasive thing it can offer enterprise AI buyers is not a smarter model but a steadier invoice. The IT services firm announced on November 10 that it will run Claude-powered systems in production on clients’ behalf under a new managed services model, which it says delivers more predictable AI spend within enterprise-grade security and governance controls. Backing the offer is a new Claude Frontier Capability Group of more than 25,000 professionals dedicated to delivering Claude for clients. For CIOs trying to move AI from pilot to production, the real question is whether handing over operations also hands over the budget risk.
Why a predictable bill is the pitch
Cognizant frames the problem as one of delivery rather than access: enterprises can reach frontier AI capability but continue to struggle to turn it into business value. Paul Smith, Anthropic’s Chief Commercial Officer, made a similar point, saying the people with the skills to make AI work inside a real business have become the hardest talent to find.
Cost is a large part of that struggle, even if the announcement does not say so directly. Consumption-based pricing is easy to approve for a pilot and much harder to forecast once agents run across thousands of workflows, and finance leaders tend to resist spend they cannot predict. By promising more predictable AI spend, Cognizant is addressing an objection that often slows scale-up, and it is doing so through a services contract rather than a technical feature.
What the managed services model covers
Under the model, Cognizant manages the operation of Claude-powered systems in production, combining dedicated talent, licensed access to Claude and its own intellectual property, including an AI Delivery Operating System. Clients get one accountable partner for running Claude in production, with each engagement run within security and governance controls customized to the client’s connectivity, architecture and risk requirements.
The bundling is what makes predictability plausible. When model access, people and tooling sit in a single arrangement, a client is no longer assembling a bill from separate usage charges, staffing costs and tooling licenses, and some of the variability moves to Cognizant.
So does accountability: CEO Ravi Kumar S described the offer as taking responsibility for delivery once AI is running inside a client’s business, a step beyond the staff-augmentation model that has long defined the sector.
What the announcement leaves out matters just as much. It gives no detail on how the pricing is structured, whether spend is fixed, capped or tiered, or how usage above expectations is handled. Predictable can mean a fixed fee, a committed range or simply better forecasting, and the difference decides who carries the risk.
The 25,000-person delivery engine behind the offer
A promise to run AI for clients only works with people to run it. The Claude Frontier Capability Group is Cognizant’s answer, a dedicated global organization building and running Claude-powered systems inside client businesses. The company says more than 44,000 associates across business units and corporate functions are trained or certified on Claude, that it holds more than 26,000 Claude certifications, and that it ranks first globally for Claude Certified Architect: Foundations.
The group will start with four priorities: redesigning business domains around AI, building and modernizing the systems that support them, operating agents securely at scale, and establishing AI-native software engineering as a core delivery capability. For modernization, Cognizant cites purpose-built Claude agents that speed analysis and refactoring of legacy and mainframe code.
Both the group and the managed services model draw on a Frontier workforce model that is producing Frontier Certified Engineers and Frontier Business Operators credentialed directly by frontier-model companies.
Cognizant offered three client examples to show the operating model at work. A global pharmaceutical manufacturer’s shop-floor agents reportedly cut downtime by 35% and shrank troubleshooting from 90 to 120 minutes down to 10 to 15. A media and entertainment company cut time to market for requirement data sheets from three to four weeks to about 15 minutes, and a North American insurer sped up campaign brief generation and retirement plan support by 40% to 60%.