What People Will Do After AI Takes Routine Work

AI recomposes roles more than it erases them. How remaining work gets done depends on engagement.

A group of CEOs at a recent meeting kept circling the same question about AI. One of them had just seen that QuickBooks is rolling out tools that can handle accounts receivable, accounts payable, and much of what an accounting manager does. 

The question in the room was the obvious one. In a couple of years, how much of this work will still be done by people?

It's a fair question. A great deal of the routine, codifiable work will move to machines, and be done well. The more useful question for a leadership team is what the remaining work becomes, and how you equip people for that shift.

Roles recompose more than they vanish

When AI absorbs the codifiable parts of a job, the work that's left changes shape. The day moves away from completing defined tasks and toward judgment under ambiguity and ownership of outcomes. A forecast still needs someone to question its assumptions. A customer still needs a person who can read a situation the model has never seen. Cross-functional work still depends on someone holding the relationships that move it forward. This is more demanding work than the routine it replaces, and it is weighted toward exactly the things software does not do.

In a 2026 Wall Street Journal survey of sixteen leading economists, including a Nobel laureate, the consensus was that AI moves work away from routine tasks and toward judgment and interpersonal skill. Ajay Agrawal of the University of Toronto called it a significant reorganization of human roles. Harvard economist Raffaella Sadun put the rising premium on judgment, persuasion, and coordination, capabilities that are inherently interpersonal. The same economists place the clear labor-market impact five to ten years out rather than two, so the direction is settled even where the timing is not.

Engagement becomes the critical variable

Here is what that shift means for staffing. Routine output tolerates disengagement. The task gets done either way. Judgment, nuance, and context are different. They require people who are genuinely invested in their work. A leaner, AI-augmented team is more dependent on each remaining person being deeply engaged and aligned to the goal.

This reframes engagement for the AI era. A company can automate the task. Caring about the result is still entirely human.

What fifteen years of data says

If engagement is now the variable that determines whether an organization gets full performance from its people, the next question is what actually drives it. CultureID has measured the same 28 questions across more than 2,000 organizational surveys and over a thousand organizations since 2012. The answers that most reliably predict a healthy, high-performing culture are consistent. People bring their best judgment and effort when they feel valued as a person, when they trust their manager, when they are clear on what good performance looks like, and when they believe the organization is fair.

The manager sits at the center of this. It is where company strategy gets translated into what an individual chooses to own on any given day. A single manager's results track the health of their whole team closely enough that, in practical terms, the manager is the culture their people experience.

The research outside our data points the same direction. A 2024 meta-analysis by Warnock, Ju, and Katz, drawing on 109 studies and more than 32,000 employees, found that trust in one's manager is the mechanism through which most relational factors translate into performance, commitment, and lower burnout. When that trust is in place, the brain treats the manager as a resource it can lean on under pressure. 

The leadership task

Let the tools take the routine work. At the same time, invest deliberately in the engagement that turns a reshaped workforce into one that owns the goals.

For most of business history, the relationship between a person and their manager was easy to deprioritize. The AI shift moves it front and center, because the work that remains for people runs on it. The relational layer can be measured, and the behaviors that strengthen it can be learned and practiced.

---

See where your own organization stands. A Company DNA snapshot shows where engagement drivers are strong across your teams, and where they are beginning to strain. Take a short tour of the assessment and the DNA Report at cultureid.com/tour.

Sources: CultureID Global Data, 2012–2026. Warnock, Ju & Katz, Journal of Business and Psychology (2024). Coan & Sbarra, Current Opinion in Psychology (2015); Rheem, Thrive by Design (2017). Economist consensus: WSJ survey of 16 economists on the future of work and AI (2026).