While many tech companies are discussing how AI will reduce the number of entry-level white-collar jobs, IT services company Cognizant offers a different perspective. CEO Ravi Kumar S. stated that companies will continue to hire a large number of recent graduates, and AI is more likely to change the structure of jobs than simply reduce the total number of positions.
Recruitment scale continues to expand this year.
At the Fortune event, Kumar stated that Cognizant hired 20,000 junior graduates last year, and this number is expected to continue to grow by 2026. Cognizant currently has over 350,000 employees and has been restructuring and laying off staff in recent years as part of its AI transformation, but the company has not stopped recruiting entry-level talent.
He believes that previous claims about the rapid disappearance of entry-level white-collar jobs are exaggerated. In his judgment, companies will still need a large number of entry-level employees and senior managers to be responsible for direction and decision-making. What will truly be reduced by AI is a portion of repetitive work in the middle levels.
New positions are not only for technical backgrounds
Cognizant recently launched its AI Builder strategy and added two new job categories: Frontier Certified Engineer and Frontier Business Operator. Kumar stated that these positions are not limited to traditional technical backgrounds.
According to him, graduates with backgrounds in history, biology, finance, and human resources may also meet the job requirements, provided they possess the ability to identify and use AI agent tools. This means that companies are placing more emphasis on tool usage and business collaboration skills in AI recruitment, rather than solely on programming or engineering backgrounds.
Oppose using token consumption to measure efficiency
Kumar publicly expressed his opposition to the current practice of many companies using token consumption to measure the intensity of AI usage and productivity. He stated that in the past two years, "how many tokens were consumed" has been more of a superficial indicator and cannot be directly equated with paid working hours, nor can it directly represent productivity.
He believes that companies should focus more on when to use AI, how to deploy AI in specific business processes, and whether it ultimately leads to verifiable business results. In other words, the focus should shift from input to output, and from project delivery and hourly billing to results-based accountability and payment.
Kumar also stated that as AI integrates into more workflows, companies' talent structures will become flatter: front-end execution and back-end verification roles will still exist, while the middle layer will be streamlined. This also means that AI's impact on employment is more akin to job restructuring than a one-way replacement.
