AI work design
Shareable reference
Before scale: who owns the KPI, who can change the workflow, who owns exceptions, and who may expand scope.
Share: https://meridian.keyteller.com/use-cases/agent-handoff-across-teams
A practical example
A support or finance pilot looks good. The director forces an explicit transfer before scale: who owns the KPI, who can change the workflow, who owns reliability, who owns upstream data, who owns policy exceptions, who may expand scope, what triggers escalation, and what triggers rollback. The build team does not keep the outcome. The process owner who can change staffing and exceptions does. Most post-pilot failure happens when a signal needs an owner with authority, not a dashboard viewer.
What you leave with. A handoff design with accept, reverse, and named owners.
Not this. A go-live party with the lab still on the pager. Calendar for this class of work is measured in weeks, not hours.
Related situations