Why AI in consulting is not replacing the human — it’s replacing the form that never was human in the first place.

So here we are. AI is breathing down our necks. We, the consultants, the organizational developers… we are attempting something strange.

Let me tell you what we actually do. We walk into organizations where people work. They are immersed in their tasks, their routines, their daily reality. And we look over their shoulders. We observe what they do, how they do it, and — most importantly — why they do it.

Like a gardener in a garden. What isn’t growing well? What is in pain? What is not what these people actually want?

Here is the paradox that still astonishes me after decades of this work: people spend years doing what they don’t want to do. And the remedy is disarmingly simple. Lead them back to what they do want. There is an inner compass built into every human being. It knows the direction. And yet, somehow, people wander away from it.

That is what we do. We help people find their way back.

Now. When we first started talking about an AI agent that could conduct interviews with employees, it caused real shock. Even among us.

Think about it. We are the people who say, loudly and often, that we stand with human beings. That our work is about helping people wake up. How could we possibly hand any part of that over to a machine?

We struggled with this. Genuinely.

But let me tell you a story first.

In 1924, researchers began studying what physical conditions make factory workers more productive. They turned up the lights. Productivity rose. They turned the lights up further. Productivity rose again. Wonderful. Science at work.

Then, to be properly scientific, they turned the lights down.

They expected productivity to drop.

It didn’t. It kept rising.

The researcher was baffled. He walked over to one of the factory workers and simply asked: how is this possible? The lights go up, you work better. The lights go down, you still work better. What is going on?

The worker shrugged. “The light doesn’t really matter,” he said. “It just feels good that someone is paying attention to us.”

This is the Hawthorne effect. And this is where our story truly begins.

When we meet an organization as consultants, the reality is almost always the same: there is not enough time, not enough money, not enough space to talk to everyone. How beautiful it would be if we could. But by the time we finished speaking with everyone, we would have to start all over again. The organization would have already changed.

So someone invented the survey.

What a brilliant idea! Hundreds, thousands, even tens of thousands of people can fill it out in a single day. Well-crafted questions. Weeks of analysis. Neat conclusions.

We did this too. Back in the early days of the internet, our first major system was exactly this — a survey platform for thousands of people, with complex, layered logic built into it. We learned a great deal from that experience.

Here is what we learned.

People will give you honest answers for about fifteen minutes. They feel that someone is paying attention. They engage. After fifteen minutes? They just want it to be over. The data collected from that point forward is worthless.

And there is more. Surveys use closed questions — because open ones are harder to process. Some questions are clear and well-formed. Others are too complex, too abstract, or simply don’t fit the context of the person answering. The participant finds them meaningless. But there is no way to stop, ask, skip, or push back. So you get a meaningless answer to a meaningless question. And you build your analysis on it.

So here is the polarity.

On one side: the survey. It can reach everyone. But it is closed, boxed, lifeless. No interaction. No real listening.

On the other side: the consultant. Deeply interactive, deeply perceptive. The consultant senses layers that no form can capture. Reads the room, follows the thread, asks the question that hasn’t been asked yet. But the consultant is extraordinarily expensive. And slow. One person at a time.

Now. Between these two, something new appears.

The AI agent does not replace the consultant. Let me say that clearly.

What it does is this: it replaces the survey. The form that was never human in the first place. The dead instrument we used because we had no better option.

The AI agent can ask open questions. It can notice when answers become flat and templated — and ask for examples. It can explain what it means. It can bring a conversation back to life when it starts to drift into autopilot. It creates a space that feels, at least partially, like someone is paying attention.

Is it a consultant? No. It cannot perceive the many layers a human being perceives. The subtle tension in a room, the thing that isn’t said, the shift in someone’s posture. That multi-layered sensing belongs to us. To the humans in this work.

But the AI agent can do what no consultant can: it can hold a thousand conversations in a single day.

So let me be precise about what is happening here.

This kind of AI-assisted conversation develops a tool that was already far less human than a consultant. The survey. The checkbox. The five-point scale. That is what is being transformed — not the human encounter.

What stays irreplaceable is this: once the conversations are gathered — the quotes, the stories, the messages from the people who spoke — it is humans who process them together. Consultants. Leaders. Team members. Sitting in a room, reading what was said, sensing what it means. Not an algorithm producing a dashboard. People, together, making meaning from other people’s words.

And through this approach, something new becomes possible. Instead of interviewing only the top leadership of a large organization and calling it a diagnosis, we can now hear from hundreds. Thousands. The richness, the texture, the sheer volume of human experience that reaches the processing table is incomparably greater.

The light doesn’t matter. But someone cares. And now, we can let more people feel that someone does.

Szabolcs Emich is Chief Innovation Officer at Atlas Platform and founder of Jövőképző, working at the intersection of organizational development, anthroposophical thinking, and technology.