The most capable
do not get relief first.

The work is finished. It is not moving forward.

The document is already finished.

The person responsible reads through what AI has assembled. The writing is polished. Nothing obviously seems to be missing. Before sharing it, they return to the beginning once more.

Can I use it as it is?

Work can stop at a question like this. AI makes production fast. What comes next—deciding what counts as good enough—remains unsettled. Read. Go back. Check. Something supposedly finished is still sitting on the desk.

On October 19, we asked KN, PROTHENTIA’s COO, about life on the ground. Before AI performance came up, the conversation turned to work getting stuck with people. AI may produce something useful while the people expected to use it cannot keep pace. Human-produced work can also be inadequate, taking time away from other priorities.

The work of producing is over. The work of deciding whether it is usable remains.

One person looks at an output and sees how it could inform management. Another cannot get past wondering whether it is correct. They are reading the same material, but looking toward different things.

Checking accuracy is necessary. But an absence of errors does not guarantee something is useful at work. What is it for? What will using it change? Without someone able to make those judgments, a polished output remains just that.

The person looks at the document again. If they knew what needed fixing, they could fix it. What they do not know is how much checking is enough to pass it on. They cannot draw the line that marks completion. So they hand the judgment, along with the work, to someone who can.

Finished, but not moving.

Before the next task, someone else’s arrives

Work that stalls in one person’s hands moves to someone who can make the call.

That person has work of their own. There are projects they want to advance. Still, they open the document, find what is missing and work out what would make it usable. Another person has spent a long time on a judgment they could make quickly.

They cannot simply ignore the gap.

They explain what needs changing and send it back. When it returns, they review it again. The specified changes have been made, but deciding whether the whole is now usable still rests with them. They can delegate the actions without being able to let go of the final judgment. This, too, is a possible pattern of handoff.

Looking at these handoffs helps explain why one person getting faster does not make the whole workplace faster. A task may be over for the person handing it off, while the recipient gains more checking and decision-making work.

We asked KN whether the people creating value seemed to be enjoying their work. They can tackle more advanced challenges, use AI, produce results and earn recognition. Surely work should have become more interesting.

The answer was not entirely upbeat.

The capable people are struggling too. They are spending time on the question of what to do about those who cannot keep up.

Fix it yourself and the immediate task moves forward. But will you have to fix the next one too? Will the person you want to trust with it be able to make the judgment next time? Making an artifact usable and helping someone become capable of owning the work remain two different tasks.

Those who can move ahead cannot simply move ahead alone.

The people who attract recognition and opportunity also attract the burden of supporting everyone around them. Time saved with AI is not automatically available for their next challenge. Other work moves into the space.

So relief does not necessarily reach the most capable first.

Opportunity brings more opportunity

Who should take on the next difficult assignment?

The name of someone who delivered results comes to mind. That person uses AI more deeply again. Through the new assignment, they learn new ways to use it and gain more experience in making judgments. Results bring the next opportunity.

Opportunity is not a one-off event.

This was the cycle that emerged from our conversation with KN. Recognition accumulates. Opportunities accumulate. Those opportunities develop capability, producing further results. Even while people outside the cycle stand still, work continues to circulate within it.

The gap is not only in what has just been produced. It is already present in what people will be trusted with next.

The roles people play are changing too. Leading value creation and transformation. Working with AI to carry out operations. Handling standardized work. These were the three roles discussed in the conversation.

These roles are not permanent labels for people. Yet, as standardized work is automated, the roles of those who did it will have to change. There is also work in supporting the people who create value. How do people move into it?

Evaluation raises another problem. Should everyone receive the same treatment? Yet we hear that treatment cannot be changed. Differences in results are visible. The way future assignments are allocated changes. Pay and conditions stay where they were.

Even those moving ahead can hesitate. A different way of working within a project can draw puzzled looks from colleagues. Fear of that reaction, KN says, can work against using AI.

A method chosen to move work forward faster can also create distance from the people around you.

More capability means more freedom. The story does not fit into that one line. What will you be entrusted with? Whom will you support? With whom will you carry out the same work? Some of the conditions that move work forward lie outside the individual using AI.

What fills the space left behind?

After a role is automated, assign its holder work involving contact with people.

It is an imaginable destination. But that person may not be good at working with others. Renaming a job does not make someone immediately able to do it.

Decide what to entrust to them. Help them acquire what they need. Have someone support them until they can do it.

All that work sits beneath a single line about reassignment. And that someone is already busy.

Assigning responsibility can be as simple as writing a name. But if work stops after that name is written, someone else must take it over again. Having someone able to take over is not the same as having a sustainable way to delegate.

The possibility of giving part of a job to AI is not the same thing as the value of the person who has done that job. Saying so, however, does not decide what to ask them to do tomorrow. We cannot leave things at affirming someone’s value while offering no work they can take on.

How should we understand what comes next? The conversation with KN did not settle that question. It extended even to systems that support people’s livelihoods, but left open whether new roles can be created and people helped to grow into them.

KN says there are executives on the client side grappling with the same issue. This can no longer be treated only as a forecast about employment in the distant future. What should the person in front of us take on next? The work of deciding is already here.

Let capable people create the next source of value. Create roles that support them. Find the next piece of work for people who have not yet entered the cycle.

All three have to move forward together.

The person from our opening scene hands over the document. The recipient opens it. There is unfinished work on that person’s desk too.

I can do it myself this time. In workplaces that have moved forward that way, who will take responsibility for what comes after?