The part that actually decides whether an AI project succeeds. The technology usually works; what fails is that nobody changed the process, the incentives still reward the old way of working, and the people expected to use it were told rather than involved. Most of the budget should probably go here.
Technical view
Established models emphasise a clear case for change, visible sponsorship, participation from affected staff in design, capability building, and reinforcement through process and incentive changes. AI-specific factors include job security concerns requiring honest communication, redesign of quality assurance around new failure modes, and revision of role definitions where the task mix has changed. Adoption should be measured by depth of use in the workflow rather than by licence count.
Related
Part of AI Terminology Index, section: Product and Adoption.