An article in MIT Sloan Management Review argues that adding AI to existing workflows does not lead to strong productivity gains.

The structure of work must be reconsidered, including which tasks people handle, which tasks automation supports, and the way work moves across humans and machines.

The authors are using as an example a financial services firm restructuring its customer order process thus achieving strong efficiency improvements after redesigning work itself.

Research across sectors supports this view and confirms larger gain when work is broken into tasks, automating routine parts, and elavating people toward higher value activities. Studies also identify workload shifts, new oversight responsibilities, and short term performance dips when teams keep old workflows.

• Redesign tasks before selecting tools.
• Plan where human capacity goes after automation.
• Build team practices that support new roles and quality control.
• Expect short term disruption, manage it proactively.

How do you plan to redesign work so AI delivers meaningful results?

Link to the article here

03 дек 2025