Ravi Kumar S. stands apart. While many technology leaders warn that artificial intelligence will hollow out junior roles, the Cognizant chief executive hires thousands of recent graduates. He calls tokenmaxxing a vanity metric. And he argues the workforce pyramid will flatten but not shrink at the base.
Kumar’s stance carries weight. He runs a company with more than 350,000 employees and roughly $27 billion in revenue. Fortune captured his views in a wide-ranging conversation at the publication’s COO Summit. There he pushed back against months of dire forecasts from OpenAI’s Sam Altman and Anthropic’s Dario Amodei. Both men later softened their predictions.
“There was a little bit of fearmongering from reading about the fact that there’s going to be a collapse of jobs,” Kumar said. “I think there will be more jobs.”
His company has felt the pressure to adapt. Cognizant pursued restructuring and layoffs under a plan called Project Leap. Yet it still brought on 20,000 entry-level college graduates last year. Kumar expects that figure to rise in 2026. The hires span traditional STEM paths and extend to history majors, biology students and even HR accountants comfortable with agentic tools.
“It could be a history major with skills to identify and use agentic work. It could be a biology major known as life sciences. It could be an HR accountant who can use agentic Claude terminals around them,” he explained.
The logic rests on a simple observation. AI handles routine middle-layer tasks. That leaves more work at the edges. Front-end roles generate prompts, interpret outputs and validate results. Back-end positions perform verification and authentication. The middle shrinks. The base expands.
Kumar’s Thesis Challenges Industry Consensus on Junior Talent
Months earlier Kumar laid out his thinking in greater detail with Business Insider. He described a “whole thesis” that more freshers, the term he prefers for recent graduates, will be needed because expertise barriers drop. AI puts specialized knowledge on fingertips. Deep technical mastery matters less. Interdisciplinary awareness matters more.
Data from inside Cognizant supports part of the claim. The bottom 50 percent of developers saw productivity rise 37 percent. The top half gained 17 percent. Lower performers benefit disproportionately. That pattern suggests organizations can absorb more juniors and still deliver results.
Kumar acknowledges uncertainty. “I don’t know what’s the right answer, but I at least know that the model is going to change,” he told the outlet. Still he bets on expansion. As AI agents multiply, engineers will spend less time writing code and more time orchestrating systems that manage those agents. Software adoption will climb because work costs less. Demand for human oversight will follow.
“More things will be reimagined, and more human labor will be needed to supplement the AI digital labor,” he said.
The message resonates beyond Cognizant. Other executives have noted similar junior gains. Amazon AI leaders and Google Cloud’s CTO have described how entry-level staff receive superpowers from these tools. Yet the broader industry narrative remains mixed. Klarna cut staff sharply after AI gains before restarting some hiring. Salesforce paused engineering recruitment. Kumar’s approach stands in contrast.
And his critique of current AI measurement practices cuts deeper. For two years companies chased token consumption as proof of progress. Leaders at Meta, Amazon and OpenAI tracked the metric internally. Dashboards rewarded volume. Employees spun up meaningless tasks to hit targets.
Kumar dismisses the practice outright. “For the last two years, how you consumed tokens, how much tokens you consumed was a vanity metric,” he said. He rejects any direct link to paid hours or productivity. Real value, he insists, must tie to business outcomes. Companies should underwrite results and get paid for them. Projects and billable hours belong to an earlier era.
Tokenmaxxing’s Rapid Fall From Favor
That view gained fresh support days before Kumar’s Fortune interview. Fortune reported that tokenmaxxing had effectively ended at many firms. The strategy produced high costs without clear return on investment. Uber burned through its annual token budget in four months. Salesforce faced a projected $300 million bill to Anthropic. Executives struggled to connect token volume to user impact or revenue.
One analysis framed the issue as a classic productivity J-curve common to general-purpose technologies. Early experimentation often reduces output before redesigned processes unlock gains. Most companies remain stuck in the trial phase. They have not yet rethought entire business lines. Tokenmaxxing simply measured activity, not progress.
Kumar’s alternative focuses on grounded execution. “It has to be grounded in the company’s hustle,” he said. Strategic token use becomes a team discipline shaped by specific workflows. Outcomes replace inputs as the yardstick. The IT services industry, long built on projects and hours, must reforge itself around accountability for results. Those who succeed will win market share.
Cognizant’s own results offer context. In the first quarter of 2026 the company posted revenue of $5.4 billion, up 5.8 percent from a year earlier. Adjusted earnings per share rose 13.8 percent. Bookings grew 21 percent, fueled by seven large deals. The performance arrived amid ongoing transformation. Project Leap targeted cost savings. AI initiatives expanded. Entry-level hiring accelerated.
Observers note the tension. A services firm that once scaled with large pyramids of billable consultants now promotes a flatter structure with more juniors at the base. Success depends on whether those juniors, equipped with AI tools, can deliver verifiable business value faster than legacy models.
Kumar remains optimistic. He sees AI as a force that widens access to expertise and accelerates paths to contribution. Liberal arts graduates and community college students gain seats at the table. Validation and verification roles multiply. The middle management layer thins but the organization grows overall.
His bet carries risk. If AI agents improve faster than expected, even entry-level oversight roles could contract. Productivity gains might not translate into higher demand. Yet Kumar’s public confidence, backed by concrete hiring numbers and internal productivity data, offers a counter-narrative at a moment when many technology leaders hedge.
The coming quarters will test the thesis. Companies will continue to experiment with measurement systems that move beyond tokens. They will watch whether expanded junior cohorts produce the outcomes Kumar predicts. For now the Cognizant leader has staked his company’s direction on humans at the edges, AI in the middle, and results as the only metric that counts.
WebProNews is an iEntry Publication