Rethinking Role Design in the AI Era

For decades, designing a job was straightforward: define a role, assign tasks, and measure outputs. That formula worked when tasks were stable and human judgment was the primary source of value. AI changes that dynamic entirely. It can handle complex analysis and generate insights at scales previously impossible, forcing organizations to rethink not just what work is done, but how it is structured.

The Great Unbundling

Too often, organizations encourage workers to use AI to speed up existing tasks. But that misses the real opportunity: reimagining roles for a world where intelligence is shared between humans and machines.

Most jobs—whether analyst, designer, or customer service rep—combine different kinds of work. Some tasks are better handled by AI, others still depend on human judgment and engagement. The challenge is to unbundle these tasks in a way that clarifies where human effort creates the most value.

Take a product team preparing a launch. Before AI, roles had clear boundaries. With AI, those boundaries blur: analysts use AI to surface patterns so they can focus on interpretation; designers refine AI-generated mockups to ensure brand consistency; product managers synthesize automated updates to concentrate on strategic decisions. The titles stay the same, but the workflow becomes a fluid partnership—allowing the team to move faster and think bigger.

That fluidity brings possibility, but also disorientation. When AI absorbs the routine, people can be left wondering what’s left for them. A blanket directive to “use AI more” without rethinking roles only fuels burnout, confusion, and wasted effort.

At its best, though, unbundling frees people from rigid job boxes and opens up a more collaborative space—one where they can lean into the work they find most meaningful, and offload what machines can handle better.

The New Organizational Blueprint

To embrace this more collaborative approach, organizations need a new kind of blueprint for role design—one that treats human and machine capabilities as evolving, interdependent parts of the same system. One way to think about this is through three core components:

1. Deconstruct the role

Before building something new, you need to map out what the role actually entails. Tasks can be grouped into three categories:

• Automation-first: Repetitive, data-heavy tasks AI can handle fully, like generating draft reports, pulling data summaries, or flagging routine anomalies.
• Augmentation-first: Tasks requiring human-AI collaboration. AI provides the foundation, but human judgment refines and contextualizes output—e.g., a designer curates AI-generated mockups, or a copywriter polishes AI text for tone and clarity.
• Human-first: Tasks relying on uniquely human skills such as emotional intelligence, mentoring, negotiation, or strategic decisions based on nuanced context.

2. Redesign the Workflow, Not the Job

With tasks deconstructed, shift from static job titles to dynamic skill clusters. A project might need "data storytelling" and "strategic judgment" rather than a "Senior Analyst." The workflow is designed with clear handoffs between human and machine, maximizing the discretionary and discerning work of the human.

3. Orchestrate with Intent

This blueprint can't succeed without a shift in leadership. The manager’s role moves from taskmaster to orchestrator. Their job is no longer to assign discrete tasks but to design a system where talent and technology can work together to create the most value. This requires leaders to cultivate a culture of learning and experimentation. The most valuable asset for any organization in the AI era remains its people – especially when they’re empowered to use AI in a way that amplifies output rather than automates their work.

The Work Ahead

As organizations reckon with AI, the real challenge isn’t just adopting new tools—it’s rethinking the philosophy of work itself. The blueprint of the past was built on fixed roles and predictable tasks. The blueprint of the future must be flexible, human-centered, and designed for orchestration. If we get this right, the outcome isn’t simply efficiency—it’s more meaningful roles, where human identity is strengthened precisely because it evolves alongside machine capability.

‍