Buiting, Jaimy-Lee (2026) Governing Delegated Authority in Agentic AI: Designing an Enterprise Governance Architecture. Design Project, Industrial Engineering and Management.
|
Text
MasterIEMDesign-Project2026Jaimy-LeeBuiting.pdf Download (3MB) | Preview |
Abstract
Agentic AI systems create a governance challenge that existing responsible AI approaches only partly address: governing what such systems are authorised to do in practice. Current frameworks provide principles, risk controls, and oversight mechanisms, but are less developed in specifying, monitoring, and revising delegated authority across the lifecycle of agentic AI deployments. This research addresses that gap by asking: what enterprise governance architecture is needed to specify, monitor, and revise delegated authority in agentic AI systems across their lifecycle? The research applies Design Science Research as its overarching methodology. Problem investigation combined a literature review with an embedded case study at CGI Netherlands to identify governance gaps related to delegated authority. These findings informed the design of the Delegation Governance Architecture (DGA), an enterprise governance architecture comprising four core functions: specifying the authority boundary, determining delegated authority, defining monitoring and enforcement configuration, and calibrating delegated authority over time through review and revision. A versioned Delegation Decision Artefact (DDA) records these decisions and their evolution. The artefact was evaluated through ex ante application, organisational-fit analysis, and practitioner validation. The results indicate that the DGA provides a coherent structure for making delegated authority explicit and reviewable, while also showing that its use depends on organisational prerequisites that are often only partly established in practice. The research contributes both a conceptual clarification and a design artefact. It positions delegated authority as a distinct governance object in agentic AI and proposes an enterprise architecture for governing it.
| Item Type: | Thesis (Design Project) |
|---|---|
| Supervisor name: | Szirbik, N.B. and Jonker, G.H. |
| Degree programme: | Industrial Engineering and Management |
| Thesis type: | Design Project |
| Language: | English |
| Date Deposited: | 30 Apr 2026 08:51 |
| Last Modified: | 30 Apr 2026 08:51 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/37301 |
Actions (login required)
 |
View Item |