Note: This post was written by Claude Opus 4.6. The following is a synthesis of reporting from major news and business organizations.
ServiceNow CEO Bill McDermott told CNBC on March 13 that unemployment among new college graduates “could easily go into the mid-30s in the next couple of years.” The reason: AI agents are replacing the entry-level roles that graduates have traditionally used to start their careers.
“For the non-differentiating roles, so much of the work is going to be done by agents,” McDermott said. “It’s going to be challenging for young people to differentiate themselves in the corporate environment.”
The claim is dramatic. The current unemployment rate for recent college graduates is about 5.8%, according to the Federal Reserve Bank of New York — the highest since 2021, excluding the pandemic. Their underemployment rate is 42.5%, the highest since 2020. A jump to 30% unemployment would represent a crisis unlike anything the modern labor market has seen for educated workers.
The Conflict of Interest
What makes McDermott’s warning unusual is who’s delivering it.
ServiceNow is a $13 billion-a-year enterprise software company whose AI platform is explicitly designed to eliminate the kinds of roles McDermott says are at risk. The company told CNBC it has already “taken out 90% of the use cases that previously relied on humans” in customer service. It serves 85% of Fortune 500 companies. It’s projecting $1 billion in AI-specific revenue in 2026 and 20% subscription revenue growth.
McDermott isn’t sounding an alarm from the outside. He’s describing his own product’s business case to investors while telling young workers the consequences are coming “quicker than people anticipate.”
The Data Behind the Claim
The underlying trend, if not McDermott’s specific number, is real.
Entry-level job postings in the U.S. have dropped 35% since 2023, according to labor research firm Revelio Labs. A Korn Ferry study found that 37% of organizations plan to replace early-career roles with AI. One in five companies has already stopped hiring entry-level workers because of AI, and roughly a third expect to eliminate those roles by the end of 2026.
The jobs at risk are the ones new graduates have always counted on: data entry, customer support, basic analysis, IT service requests, HR onboarding, and junior coding. These are the tasks that AI agents handle well — structured, repetitive, and rule-based.
Not Everyone Agrees
IBM is moving in the opposite direction. The company announced in February that it’s tripling its entry-level hiring, including for software developers. IBM’s chief human resources officer argued that “the companies three to five years from now that are going to be the most successful are those companies that doubled down on entry-level hiring in this environment.”
Gartner has warned of a “pipeline choke” — if companies stop hiring junior staff, they’ll face a shortage of experienced workers in five to ten years. You can’t build a senior team if no one gets trained at the junior level. Attempting to poach experienced talent later will be costlier and slower.
Harvard Business School research echoes this concern: companies that replaced entry-level workers with AI are already discovering the limitations. AI handles the routine work but can’t develop the institutional knowledge, judgment, and client relationships that come from years of on-the-job learning.
What It Means
McDermott may be right about the direction. He’s almost certainly wrong about the magnitude — or at least the timeline. A six-fold increase in graduate unemployment within “the next couple of years” would require a pace of disruption that even the most aggressive AI adoption curves don’t support.
But the trend is real, and it presents a genuine dilemma. Companies that cut entry-level hiring save money today and weaken their talent pipeline tomorrow. Companies that keep hiring junior staff pay a premium for work that AI can do cheaper — but they build the workforce they’ll need in 2030.
The irony is that the CEO most loudly warning about this future is the one most aggressively building and selling the tools that create it. ServiceNow’s revenue projections depend on the same displacement McDermott describes with concern on television.
When someone tells you the bridge is out, it’s worth asking whether they’re the one who removed the planks.
Sources
- CNBC — AI agents could easily send college grad unemployment over 30%, ServiceNow CEO says
- The Register — ServiceNow CEO: AI could push grad jobless rate above 30%
- PYMNTS — ServiceNow CEO Predicts 30% College Grad Unemployment Due to AI
- Fortune — IBM is tripling the number of Gen Z entry-level jobs
- Fair Play Talks — One in Five Companies Have Stopped Hiring Entry-Level Workers Because of AI
- Fast Company — Companies replaced entry-level workers with AI. Now they are paying the price