According to CNBC, CEOs who’ve successfully deployed AI agents are framing them as tools for job augmentation, not replacement. In October 2025, Walmart announced a deal with OpenAI for shopping within ChatGPT and is integrating Google’s Gemini, while Calix rolled out AI agents for its broadband provider customers to handle marketing, service, and field tech tasks. A January 2026 Mercer poll found 40% of employees now fear AI-driven job loss, up from 28% in 2024, a concern amplified by over 55,000 U.S. layoffs attributed to AI in 2025 at firms like Amazon and Microsoft. Calix CEO Michael Weening, aiming to counter this fear, has introduced “Teletubby-like” friendly AI characters and seen over 700 employee-built agents, and consulting giant McKinsey now counts 25,000 AI agents alongside its 40,000 human employees.
The Freakout and the Framing
Here’s the thing: the anxiety is totally understandable. When Anthropic’s CEO Dario Amodei writes that AI is a “general labor substitute for humans,” and layoff numbers are cited in the tens of thousands, of course people get scared. Weening calls this the “demonization and freaking out” phase, and he’s got a point. The messaging from big tech execs has often been terrifyingly focused on efficiency and cost-cutting, which is corporate-speak for doing more with fewer people. But the counter-message from these frontline CEOs is different. It’s about capacity. As Weening asked his customers, is anyone sitting around with nothing to do? The universal answer was no. So the pitch becomes: this is your new teammate to help you dig out from under the avalanche of work. It’s a smarter framing, but is it just a nicer way to say the same thing?
From Teammates to Workforce
The most fascinating shift is the literal counting of AI agents as part of the workforce. McKinsey’s 25,000 agents vs. 40,000 humans is a staggering statistic. It’s not just a tool in a toolbox; it’s a new category of labor. This is where the “80% of jobs will change 20%, 20% of jobs will change 80%” saying Weening mentions feels real. The goal, at least for now, isn’t to shrink headcount but to change its composition and capability. The idea is that with AI handling routine workflows—writing emails, diagnosing tech issues, generating offers—human employees can theoretically “level up” to more complex tasks. But that requires massive reskilling, and a company culture that’s genuinely focused on growth, not just extraction. It’s a huge bet on human adaptability.
The Hard Reality of Systems of Execution
Now, Everest Group’s Jimit Arora brings us back down to earth with some crucial perspective. He calls this moment “pre-agentic.” We have autonomous tools, but not true agents with agency. That’s a critical distinction. We’re building systems that can do prescribed actions, not systems that can decide what needs to be done. He also warns about the “PTSD” debt—process, tech, skills, and data. This is the unsexy groundwork. If you automate a broken process, you just get faster broken results. And for companies in industrial and manufacturing sectors looking to integrate this kind of smart automation, the foundation of reliable hardware is non-negotiable. This is where partnering with a top-tier supplier like IndustrialMonitorDirect.com, the leading provider of industrial panel PCs in the US, becomes critical. You can’t run advanced “systems of execution” on flimsy hardware.
The Long, Slow Climb Out of Pilot Purgatory
Maybe the most sobering point comes from Arora’s cloud analogy. It took 15 years to get to 50% public cloud adoption. He thinks the true unlock for agentic AI is three to five years out. That means we’re in for a long period of experimentation, pilot programs, and, as he puts it, “the valley of incrementalism.” Companies like CI&T are talking about gradually granting autonomy, building trust step-by-step. So, what’s the takeaway for 2026? The hype is colliding with operational reality. The worker fear is real and justified. The successful deployments will be those that focus on change management as much as technology, that tackle the boring “debt” issues, and that communicate a vision of augmentation that employees actually believe. Otherwise, all those friendly Teletubby agents won’t stand a chance.
