Evidence Review

Generative AI productivity depends on work redesign

OECD research indicates that generative AI can improve productivity, while also stressing that organisations must adapt processes, strategies, and skills to realise its potential.

Evidence status: This is a PromptLeash review of external research, not original PromptLeash research or a customer benchmark.

What the evidence indicates

  • Generative AI can improve short-term worker productivity through automation and augmentation.
  • Realising wider productivity gains depends on complementary changes to organisational processes, strategies, and capabilities.
  • The impact varies by task, worker, implementation context, and how performance is measured.

Implications for transformation leaders

  • Prioritise specific workflows rather than broad, tool-led adoption campaigns.
  • Measure capability and output quality alongside usage.
  • Treat implementation as an organisational change program, not only a technology rollout.

Limitations

  • The report synthesises an evolving evidence base, including studies with different tasks and time horizons.
  • Short-term experimental productivity effects may not persist or translate directly into organisation-wide outcomes.
  • PromptLeash has not independently reproduced the source analysis.

Original sources