Reading paths
AI adoption data
Compare business use, worker use and organizational surveys without mixing denominators.
Explore this topic →Research on performance
Inspect the task-specific experiment and workplace productivity evidence.
Explore this topic →Technology choices
Understand retrieval, models, automation and integration requirements.
Explore this topic →Governance and security
Connect responsibilities, permissions and evaluation to an operating workflow.
Explore this topic →Costs and measurement
Separate capacity value from cash savings and examine total cost.
Explore this topic →Consultant selection
Use the same evidence-based questions for every candidate.
Explore this topic →
How to read the sources
Government surveys describe specific populations and periods. Experiments test particular tasks under stated conditions. Technical guidance explains methods and risks. None alone establishes that a particular consultant will produce a particular return.
The source directory below matches the main article. The figures are original redrawings of reported data. The downloadable chart data CSV preserves the values, units and populations used in the graphics.
Research versions and updates
Use the same research version when comparing results. The consulting experiment is summarized using Harvard Business School’s April 2026 account of the published study. The worker-use figures incorporate the Census Bureau’s September 2026 weighting revision. Publication dates and observation periods remain distinct.
The editorial policy explains how recommendations and illustrative examples are labeled. Substantive revisions belong in the history record.