Why I think the"AI Job Apocalypse" is the Wrong Conversation

McKinsey recently found that 57% of work hours in the United States are technically automatable right now. That number is massive. It is easy to see why it creates panic in boardrooms and breakrooms alike.

Goldman Sachs research tells a bit of a different story. While a quarter of our daily work could be handled by AI, only 7% of entire jobs are candidates for complete substitution. That gap between 57% and 7% is significant. And the reason it exists is simple: we are using the wrong unit of analysis when we talk about AI and work. AI does not eliminate jobs. It automates tasks within jobs. Until we internalize that distinction, we will continue making bad decisions about our people, our organizations, and our technology investments.

A job is not one thing. It is a bundle of tasks.

Think about what a job actually is. It is not a single, indivisible thing. It is a bundle of tasks. Most professional roles are made up of 20 to 50 distinct activities. Some of those tasks are highly codified, repetitive, and structured. Others require deep context, relationship building, and nuanced judgment.

When we deploy technology, it does not interface with the job as a whole. The unit of automation is the task. Algorithms are incredibly good at the codified parts of our work. Pulling data, summarizing meeting notes, generating first drafts, routing standard requests. But they cannot repair trust with an angry client, navigate the unwritten political dynamics of a cross-functional project, or provide the human accountability required when things go wrong.

There is a category of work I call "glue work." It is the invisible stuff that holds organizations together. Context switching between stakeholders. Reading the room in a tense meeting. Knowing which executive needs a heads up before a decision gets escalated. None of this shows up in an efficiency metric, but remove it and the whole system breaks down.

The jobs that survive and thrive are the ones where humans handle this complex, unstructured work while AI takes care of the routine substrate. Clinical radiologists are a perfect example. They operate in one of the most AI exposed fields in medicine. Yet their employment grew by 3% per year between 2017 and 2024. Because interpreting an image is just one task in a radiologist's bundle. They also consult with attending physicians, perform physical procedures, and explain complex diagnoses to patients. The AI handles the initial scan analysis, freeing the radiologist to focus on the higher value, human centric parts of the role.

What 135 agents taught me about augmentation.

We are living this at Vasion right now. We currently run 135 AI agents across our organization. These agents are actively integrated into our workflows, handling defined scopes of work within our Digital Worker 7 framework. Not one person has been replaced by these systems. Not one.

When I map these agents on our Agent Maturity Grid, most are still operating at the "Intern" level. They are excellent at narrow, well defined tasks, but they require human oversight, human context, and human judgment to be effective. They augment our teams. They handle the repetitive work so our people can focus on the problems that actually require a human brain.

The goal was never to replace people. The goal was to free them from the tasks that were draining their energy and limiting their impact.

The Klarna cautionary tale.

The danger comes when leaders confuse task automation with job elimination and push for aggressive headcount reductions. We saw this play out publicly with Klarna. They replaced 700 customer support roles with an AI assistant. The immediate cost savings were obvious and the market rewarded them for it.

But the glue work that those human agents performed was lost. The empathy, the contextual problem solving, the trust repair with frustrated customers. Customer satisfaction collapsed. The company had to reverse course, and the CEO publicly admitted they went too far.

This is what happens with "so-so automation." You replace humans without actually improving the underlying process or the customer experience. You cut costs on paper, but you create brittleness in your operations. When you strip away the people who provide the glue work, the system fractures under stress.

Eighty-eight percent of organizations are now using AI for at least one business function. The question is not whether to adopt. The question is whether you are automating to elevate your people or simply automating to eliminate them.

The broken career ladder.

The fear of senior professionals losing their jobs is largely overblown. But there is a very real crisis emerging at the other end of the career spectrum. The risk is not that experienced workers will be replaced. The risk is that the entry level on ramps are disappearing.

The white collar career model has always involved hiring recent graduates to execute routine, codified tasks. While doing this routine work, they observed senior professionals and gradually developed the tacit knowledge required to solve complex, unstructured problems. That is how expertise gets built. You learn by doing the simple stuff alongside people who do the hard stuff.

By automating this routine substrate, organizations are removing the primary developmental pipeline for the next generation. High frequency payroll data shows a 16% relative decline in early career employment in AI exposed fields. Twenty-one percent of companies have already frozen entry level hiring due to AI, with another 15% expecting to do so by the end of 2026.

Meanwhile, experienced workers aged 45 and older have seen employment growth of 6% to 9% in those same AI exposed fields. The AI complements their tacit knowledge, making them more productive and more valuable. The technology rewards experience.

This demographic split is the true structural challenge. If we automate away the entry level tasks, how do we train the next generation of experts? Where do the senior professionals of 2035 come from if they cannot get their start in 2026?

The false narrative of "AI layoffs."

There is also a troubling trend where business leaders use AI as a convenient scapegoat for broader financial pressures. Nearly 60% of leaders admit to framing layoffs as "AI driven" when the underlying driver is actually a need to cut costs or reallocate capital toward infrastructure investments. This creates a false narrative that AI is a job killer, when the reality is more nuanced. Job openings still sit at 7.594 million. The labor market remains tight. The economy is not collapsing under the weight of automation.

Elevate before you eliminate.

History is on our side here. Sixty percent of workers in 2022 were in jobs that did not exist in 1940. Technology has always displaced some tasks while creating entirely new categories of work. The displacement and reinstatement cycle is not new. What is new is the speed at which it is happening, and the intentionality required to manage it well.

When an AI agent takes over the routine tasks within a job bundle, the goal should not be to eliminate the role. The goal should be to expand the human decision rights and elevate the worker to focus on the tasks that require judgment, context, and relationships. We must also deliberately design new pathways for entry level workers. If the routine tasks are gone, we need new models of apprenticeship that focus on developing tacit knowledge and complex problem solving from day one.

AI is not here to take your job. It is here to take your tasks. The question is what we choose to do with the time and capacity it gives back.

How are you redesigning the roles on your team to ensure your people are elevated by AI rather than competing with it?