College Grads Are Anxious About AI. Here’s How Operations Leaders Can Help
Leaders can turn campus anxiety into adoption momentum with small-scale initiatives like project-based work and updating internship roles.
Former Google CEO Eric Schmidt was repeatedly and loudly booed by students at a recent University of Arizona graduation. He wasn’t spouting AGI doom, nor floating utopian hyperbole. He was just offering level-headed advice on how to think about working with AI agents in students’ future careers.
He barely got through the speech.
AI Anxiety Is a Threat to Plans for Agentic Systems
Schmidt was not the first person booed for pitching an AI-heavy future. He won’t be the last. In fact, with growing local resistance to data centers and worries about a permanent underclass, it seems the K-shaped economy is real. AI is becoming a divisive topic on political agendas around the world, which means technical standards, regulations, and costs are up for grabs.
Meanwhile, CEO expectations of AI productivity gains put a huge change management burden on operating leaders seeking 20-40% headcount reductions with the very agentic systems Schmidt is talking about. Zero100 surveyed 100 COOs in Q4 2025 and found 83% believe less than half of their CEO’s AI commitments are feasible within stated timeframes. 42% also believe AI will increase the risk of labor strife as an external shock to the business.
Making matters worse, a new survey from Oliver Wyman shows CEOs decisively shifting away from young talent toward mid-level roles over the next couple of years. The deck is stacked against new college graduates, and their anxiety makes sense.
Plus, the anxiety is contagious. I have spoken off the record to planning and sourcing leaders worried about training the very agents that will replace them. One chief procurement officer leading a big agentic AI system rollout said that “having AI cynics in your business can derail a transformation.”
For people with college-age kids, it’s a double whammy. Everyone feels like they’re playing a high-stakes game of musical chairs.
How can supply chain leaders help?
Start with the Human-in-the-Loop
Zero100 research on agentic systems is grounded in business process. We’ve found the biggest payback comes from connecting tasks in cross-functional workflows which combine human and agent workers. (We call these PowerThreads.) The design principle of keeping humans-in-the-loop is central not only to training and scaling agentic systems, but also to meeting their specific cybersecurity requirements.
To make this design principle practical, we have broken down supply chain work at a granular level to understand which jobs will be led by humans, which will be led by machines, and which will be shared along a spectrum ending in autonomous systems. The role of humans isn’t disappearing; it’s changing over time.
This phased transition includes specific “jobs to be done” within larger roles like, say, logistics specialist. Many of these are a good fit for college interns helping mid-level managers navigate the introduction and scaling of agents within a workflow.
In one anonymized example constructed for an F500 CPG company, several such jobs to be done, including supplier coordination, load sequencing, and product launch planning, are still human-owned with AI support. These areas could benefit from having a freshly educated, technology-savvy 19-year-old around for a while. Veteran Manager gets practice leading change while Newbie Intern sees light at the end of the tunnel. Senior leadership moves faster and with less resistance to its agentic future.
Partner with the Young to Open a Few Doors
Most discussion of how to ease the transition to an AI future focuses on mega-stories like social safety nets, or Schumpeterian creative destruction generating new kinds of work. Maybe the AI titans or politicians will come to our rescue, but perhaps the humbler move is better – countless small-scale initiatives by operations leaders building agentic systems to offset fear with opportunities for the young.
The path to do so is not complicated. Review your organization’s existing internships and training tools to see if they’re up to date. Consider upskilling programs with a bias toward entry-level roles like Multiverse, a UK-based AI training platform.
And don’t expect to add entry-level headcount for career-track jobs while cutting it elsewhere. That’s not the future anyway. Instead, think project work, maybe virtual, with limited timeframes, possibly earning course credit instead of pay.
The University of Tennessee, Arizona State, and MIT, for instance, have a series of student engagement models that don’t cost much but do foster a contagious learning dynamic.
Much better than a chorus of boos.