The biggest challenge of Agentic AI isn’t AI; It's redesigning work

Most boardroom discussions about agentic AI focus on what AI can do. Our research suggests a more important question. What happens when work itself changes?

Multiple conversations. One insight.

Through six conversations conducted by the Future Makers Research Community in collaboration with Radboud University, one finding stood out consistently. The most significant impact of agentic AI is not technological. It is organizational.

The research brought together the Atos Group's industry perspective and academic expertise on the future of work. In an earlier article, we had explored key insights from those conversations. This article focuses on one conclusion that emerged particularly clearly: the real transformation is about people, work and organizational design, not technology alone.

For decades, digital transformation largely meant making existing processes faster, cheaper and more efficient. Technology automated repetitive tasks while employees continued working within largely unchanged operating models. Agentic AI represents a fundamentally different shift.

Previous waves of automation changed tasks. Agentic AI has the potential to change entire workflows. Rather than simply supporting employees, AI agents increasingly participate in workflows by analyzing information, coordinating activities and initiating actions. This means organizations can no longer simply ask what AI can do; they must also ask how work itself should be redesigned. That distinction is more significant than it may first appear.

Change is the only constant

Many discussions around AI continue to focus on automation and productivity. Yet organizations rarely create lasting value by automating isolated tasks.

The greatest opportunities emerge when entire workflows are redesigned around the complementary strengths of humans and intelligent agents. This changes the role of almost every function within the enterprise.

For executives, the question shifts from technology adoption toward organizational design. Which decisions should remain human? Where can AI operate autonomously? How should accountability, governance and leadership evolve? These questions highlight an important reality: Agentic AI is not simply another technology implementation. It changes the operating model of the organization itself.

Expertise development: The widening gap

Perhaps the most overlooked challenge of agentic AI is expertise development.
Traditionally, professionals developed expertise by performing operational work before gradually moving into more strategic roles. As AI performs more of these routine activities, organizations face a new challenge. How can future professionals develop the judgment required to supervise intelligent systems if they no longer gain experience through those operational tasks?

Addressing this question requires organizations to rethink learning pathways, career development and knowledge transfer, not only the technologies they deploy.

This is why redesigning work becomes just as important as deploying AI.

Three focus areas for agentic enterprises

If agentic AI is changing how work is performed, organizations need a practical framework for responding. Based on the themes emerging from our research, three areas deserve particular attention.

Roles - Organizations must determine which activities remain primarily human and which can be delegated to intelligent agents. As AI takes on more operational and analytical tasks, human contribution increasingly centers on judgment, creativity, relationship management, ethical reasoning and decision-making.
Workflows - The greatest value rarely comes from automating individual tasks. It comes from redesigning end-to-end workflows around the complementary strengths of people and intelligent agents. This requires new approaches to collaboration, governance, accountability and oversight.
Learning - Perhaps the most overlooked challenge is expertise development. For decades, professionals have built judgment through experience. If intelligent agents increasingly perform the work that once served as a training ground, organizations must rethink how expertise is developed, transferred and sustained for future generations of employees.

Organizations that address all three dimensions together will be better positioned to create sustainable value from agentic AI than those that focus on technology deployment alone. The organizations that succeed in the agentic era will therefore not necessarily be those with the most advanced AI capabilities. They will be those that redesign work most effectively, enabling people and intelligent agents to complement one another rather than compete.

Success in the agentic era will depend less on AI capabilities and more on the ability to redesign work.

Redesigning work for a sustainable enterprise

Ultimately, agentic AI should not be viewed as the next step in automation. It should be viewed as the catalyst for redesigning how work is organized. Technology may initiate that transformation, but people, leadership and organizational design will determine whether it succeeds.

Agentic AI challenges organizations to rethink how people learn, collaborate and create value. The question is no longer whether organizations will adopt agentic AI. The question is whether they are prepared to redesign work around it.

This article is based on research conducted by Bregje Seising, an MSc Strategic Management graduate from Radboud University who completed a research internship with the Atos Future Makers Research Community (FMRC). The article was reviewed by Petra Hofstett and Martin Pfeil. More information about the Future Makers Research Community is available here.