Organizational Redesign (Brynjolfsson)

In the AI Story

The scope of redesign extends across multiple organizational dimensions. Teams need restructuring as the skills that define specialist boundaries shift. Hierarchies need flattening as the information-aggregation function of middle management gives way to AI-augmented direct access to insight. Decision-making processes need acceleration to match the speed at which AI enables action. Performance metrics need revision as the outputs worth rewarding shift from execution quality to judgment quality. Coordination mechanisms need updating as the friction of cross-specialty collaboration drops toward zero.

The obstacles to redesign are formidable. Existing team structures represent accumulated organizational capital — people who know each other, workflows that have been debugged over years, coordination patterns that are invisible to outsiders. Tearing this down to rebuild is costly, disruptive, and often politically contested. The people with the most detailed knowledge of existing processes often have the strongest interests in maintaining them. The people proposing redesigns often lack the practical understanding of why existing processes work the way they do.

The recursive quality of AI-driven redesign is novel and consequential. AI can help organizations understand how to reorganize — simulating different structures, facilitating communication across restructuring efforts, supporting the training that redesigned work requires. The technology is not merely the object of the redesign. It is a potential instrument of it. This creates both opportunity and danger. The opportunity: organizational adaptation can be accelerated by the technology driving the need for adaptation. The danger: organizations might use AI to design their own reorganization without the human judgment necessary to ensure the redesign serves all stakeholders, not merely efficiency.

The practical prescription Brynjolfsson has advocated is publicly funded organizational innovation centers — institutions that would study, document, and disseminate best practices in AI-augmented work design. The model has precedent: agricultural extension services performed a similar function during the mechanization of farming, translating knowledge generated by research institutions into practical guidance individual farms could implement. For AI, the equivalent would be research laboratories combined with technical assistance programs, helping small and medium-sized firms undertake redesigns they lack the resources or expertise to implement independently.

Origin

Organizational redesign as a complementary investment theme runs through Brynjolfsson's work from his earliest empirical papers, but it was developed most systematically in his work with Lorin Hitt, particularly the 2000 Journal of Economic Perspectives paper Beyond Computation and the 2003 paper Computing Productivity: Firm-Level Evidence.

The AI-specific application was elaborated in Brynjolfsson's 2019 HBR article with Andrew McAfee on how AI would transform business and in his 2023 work on productivity measurement with Mu-Jeung Yang and others.

Key Ideas

Primary determinant of AI outcomes. The variance in AI productivity gains across firms is explained primarily by organizational redesign, not technology selection.

Multi-dimensional transformation. Teams, hierarchies, processes, metrics, and coordination mechanisms must all change together.

Existing structures are high-friction obstacles. Accumulated organizational capital resists reorganization even when reorganization would produce greater value.

Recursive potential. AI can accelerate its own enabling redesigns — with both opportunity and danger.

Policy support is needed at scale. Small and medium-sized firms require institutional assistance equivalent to agricultural extension services.