What happens to the people?

The question nobody asks on the first call—but everyone is thinking about—is the one about the people.

What happens to the jobs? If AI can do all of this, what happens to the team?

After two years inside companies deploying this, we have a real answer. It’s honest, and parts of it are uncomfortable.

The uncomfortable truth

Most of the work humans do today, they do by default. Not because they’re the best option for the task. Because there was no alternative.

Drafting emails, formatting reports, following up on open items. Reconciling data between systems. Logging the same information into three different places.

None of this requires human judgment. It requires human attention and human time. And there’s a limited supply of both.

AI doesn’t run out of attention or forget to follow up. It doesn’t hit 4pm and lose focus. It doesn’t have a hundred open files and quietly lose track of which ones matter.

That’s the uncomfortable part. A large percentage of what knowledge workers do every day can be done by a well-trained agent. We’ve seen it in real deployments. The percentage is significant.

What we saw in the field

At one company we work with, coordinators manage over a hundred open files at any given time. They’re fielding calls, scheduling appointments, chasing insurance adjusters, following up with homeowners, and entering data into three different systems.

They are good at their jobs. They work hard. And they are drowning.

They can’t keep up. Not because they lack effort or skill, but because the volume of follow-ups, data entry, and coordination exceeds what a human can reliably track. Balls drop. Not because anyone is careless. Because there are too many balls.

When we deployed the AI, something changed. The system started handling the follow-ups—texts, emails, status checks. It started prioritizing the to-do list based on urgency and SLA deadlines. It started flagging the files that needed human attention versus the ones that just needed a ping sent.

The coordinators didn’t lose their jobs. They stopped drowning.

Redistribution

The word “replacement” misses what’s actually happening. It’s a redistribution.

The work humans did because there was no alternative is getting redistributed to systems that do it faster, with more consistency, at scale. What’s left for the human is the work that actually requires being human.

Judgment. Relationships. The ability to read a room and know when to push versus when to wait. Understanding nuance that doesn’t fit in a knowledge file.

At the same company, the coordinators who used to spend four hours a day on follow-ups now spend that time talking to customers. Having real conversations. Building relationships. Solving problems that need a person.

They went from data entry with occasional human interaction to human interaction with AI-managed data. Same role. Fundamentally different job.

The organizational shift

The traditional org chart is built around human capacity. You need X people for Y volume. Volume grows, so you hire. More people means more managers. More managers means more layers. The organization gets heavier.

AI changes the math. Volume still grows, but it doesn’t require proportional headcount. One person with three well-trained agents can handle what used to require a small team. A department head orchestrating a group of agents can move work that used to require an entire department.

This doesn’t mean companies shrink to five people overnight. It means the ratio shifts. The number of people who think—who make decisions, set strategy, manage relationships—stays roughly the same. The number of people executing repetitive, process-driven tasks shrinks as those tasks get absorbed by agents.

Some companies will use this to reduce headcount. Others will use it to dramatically increase output with the same team. Most will do a combination.

The timeline

We’ve already seen digital jobs change in real time. The coordinators, the analysts, the compliance reviewers—they’re working differently today than they were a year ago. Anything that involves moving information between systems, formatting documents, or managing communications is being transformed right now. We’re in the middle of it.

Knowledge work comes next—analysis, reporting, content creation. By 2028, AI support in these functions won’t be an advantage. It’ll be the expectation. Physical work waits for robotics and humanoids, which are progressing fast but aren’t there yet.

The pattern is gradual, and then all at once. The companies preparing now won’t be caught off guard.

The most valuable skill

We keep coming back to this.

When we’re training a team to work with AI, we always say the same thing: the most valuable skill in the next five years is understanding how to use AI in your day-to-day work. Not building it. Not coding it. Using it. Knowing how to frame a task. Knowing how to review an output. Knowing when to trust the system and when to override it.

The people who develop this skill become the most valuable person in their department.

One person who understands how to work with agents can do the work of three or four people who don’t. That person becomes the hub of the department. The one everyone goes to. The one who can’t be replaced because they’ve become the interface between human judgment and machine execution.

The career move nobody is talking about enough: become the person who knows how to work with the system.

People think. AI executes.

Strategy, judgment, relationships, creativity—that stays human. The drafting, processing, formatting, follow-ups, and reconciliation—that goes to AI.

The organizations that get this wrong put AI in the thinking seat and humans in the execution seat. They get bad strategy with expensive execution. The ones that get it right free their people to operate at a higher level while the AI handles the volume.

The goal isn’t fewer people. The goal is people doing the work only people can do.

What we tell leaders

When a leader asks us what’s coming, we don’t sugarcoat it. A lot of routine work is going away. The question isn’t whether. It’s when and how fast.

But we also don’t catastrophize. What we’ve seen in practice is more nuanced than the headlines suggest. People aren’t getting replaced by AI. They’re getting amplified by it. The coordinators who were drowning are now managing relationships. The analysts who spent three days on a report now spend three hours and the rest of the week on the decisions the report informs.

Build the knowledge base while your best people are still here to contribute to it. Train the team to work alongside the system. Design it so that the human work—the work that’s left—is the meaningful work.

The water level is rising. It’s been rising for two years. But the companies we work with aren’t drowning.

They’re building boats.