A business unit wants to add a document source to an existing AI search tool. Who checks data access, who approves the change and who rolls it back if results are wrong? Develop a change procedure that assigns these tasks and handovers.
Trace the request through two organisational models: a central decision and a trained business role deciding within shared rules. Compare responsibilities and turnaround times, review a work sample and establish rollback. If the source contains sensitive data, apply the escalation route before implementation.
Create a change procedure with roles and handovers for your own area or the case. Reuse it for further AI changes and discuss it with the people authorised to decide.
What you will work on
Justify central and business-unit tasks for an AI use case
Assign decision rights, set minimum standards and allocate exception authority
Trace the same change through two organisational arrangements
Who it is for
COOs and organisational leaders
AI leaders
Organisational development professionals
Topics
Scope of the AI operating model
Central and business-unit roles
Decisions across the application lifecycle
Minimum standards and exceptions
Handovers and organisational adaptation
Your result
Your change procedure records central tasks, business-unit contributions, handovers, exceptions and feedback. Use it for further AI changes and discuss it with the people authorised to decide.
Prerequisites
Knowledge of organisational processes and basic AI applications. A defined organisational area with sufficient insight into roles and processes, or a case study, is sufficient.
Format and approach
Planned: 8–10 hours of self-study, with organisational design tasks and self-checks with explanations.
An AI tool is meant to sort incoming requests. Employees demonstrate that urgent cases are classified as routine and no timely second review is in place. Assess their evidence and change the workflow before planning further rollout.
Design a process for AI incidents, from reporting and limiting use through business review and a possible restart. Follow up decisions already influenced by an incorrect output.
Together, set up an AI team for an organisational process. Allocate tasks to roles, design handovers and establish human decision rights. Test and improve the workflow using agreed cases to inform a decision on further use.