Redesigning the Work Means Redesigning the Roles

The conversation about AI and workflow redesign often starts in the right place but stops too early.

McKinsey's State of AI research found that workflow redesign has one of the strongest relationships with bottom-line impact from generative AI and that organizations seeing stronger AI performance are far more likely to redesign how work gets done. That finding is important because it confirms what many leaders are beginning to see: meaningful value does not come from inserting AI into yesterday's process. It comes from redesigning how the work should happen.

But I think there is a second implication that deserves more attention.

You cannot fundamentally redesign a workflow without fundamentally changing someone's job.

A workflow is not just a process diagram. It is someone's day. It is where judgment is applied, where decisions are made, where accountability sits, where handoffs occur, and where people understand the value they contribute. When AI changes the flow of work, it also changes the shape of roles. Some tasks disappear. Some are augmented. Some decisions move closer to the customer. Some roles become more strategic, while others become narrower, more analytical, more relational, or more dependent on new forms of human-machine collaboration.

This is where many organizations create avoidable anxiety. Workflow redesign is often owned by process, technology, or transformation teams, while role design sits in a different conversation. The process team maps the future state. The technology team evaluates tools. Leaders communicate the business case. But the people currently doing the work are left to interpret what the change means for their responsibilities, authority, skills, and future.

That gap is not harmless. Ambiguity about whether a role is being augmented or absorbed is corrosive to trust, engagement, and participation. People can adapt to change when leaders are clear about what is changing and why. What they struggle with is silence, especially when the silence surrounds their own work.

The organizations that handle this well do three things earlier than most.

First, they name which work is changing or disappearing instead of allowing employees to infer it. Second, they define where human judgment will continue to govern, not as a limitation on technology but as a deliberate design choice about where discernment, accountability, and relationship still carry value. Third, they build capability toward the redesigned work, rather than training people to use a tool inside a job that is already being reshaped.

That is the people side of AI strategy, and it cannot be treated as an afterthought.

Leaders need to ask practical questions before the rollout is underway. What work should no longer exist? What work should be augmented? Where must human judgment remain explicit? What new capabilities will the redesigned role require? Which old measures of performance no longer make sense? What will employees need to stop doing, start doing, and learn to do differently?

The leadership question I would put to any team in the middle of workflow redesign is this: can you describe what the redesigned roles look like - and has anyone told the people currently in them?

That is where AI transformation becomes real. Redesigning the work is the strategy. Redesigning the roles is the delivery.