Conference item icon

Conference item : Abstract

Becoming an agentic enterprise: a practitioner methodology and frameworks for human-AI governance

Abstract:

The transition from pilot projects to enterprise-scale autonomous AI systems represents one of the most significant governance challenges facing organizations today. While much attention has focused on the technical capabilities of large language models and agentic systems, comparatively little research has addressed the practical governance frameworks and methodology required to manage portfolios of agents at scale within an Enterprise. This creates a governance gap that keeps enterprises in long pilot cycles. For enterprises to confidently deploy, they need governing frameworks: Boards lack comprehensive frameworks to assess, classify, and oversee these increasingly sophisticated human-algorithm systems. Transformation leaders need a way to assess the human oversight required as agent complexity evolves.


This paper introduces the “Becoming an Agentic Enterprise” methodology developed by the author (s) as a comprehensive practitioner framework, grounded in direct engagement with enterprise transformation initiatives. The methodology comprises four interlocking components: SCORE-AI for board-level governance assessment, the L1-L6 Agent Maturity Model for capability classification, the AURA framework for human oversight calibration & the extent of integration required with existing systems and processes, and the RRR transformation sequence for phased implementation. We illustrate the application of this integrated methodology through a detailed case study of an enterprise software company that uses this framework, resulting in a blueprint for ~100 agents across eight business domains spanning the spectrum of agent maturity. The paper contributes to the human-AI interaction literature by demonstrating how layered governance structures can maintain meaningful human oversight while enabling operational autonomy, and by providing a replicable blueprint for enterprise-scale agentic transformation.
Publication status:
Published
Peer review status:
Reviewed (other)

Actions

Access Document

Files:
Publisher copy:
10.2139/ssrn.7123579

Authors

More by this author
Institution:
University of Oxford
Division:
ContEd
Department:
Oxford Lifelong Learning
Oxford college:
Jesus College
Role:
Author
ORCID:
0009-0004-6360-188X


Publisher:
Elsevier
Journal:
SSRN More from this journal
Publication date:
2026-07-16
Acceptance date:
2026-07-06
Event title:
Human-algorithm interaction workshop 2026
Event location:
Oxford, UK
Event website:
https://www.sbs.ox.ac.uk/events/human-algorithm-interaction-workshop-2026
Event start date:
2026-07-06
Event end date:
2026-07-08
DOI:
EISSN:
1556-5068

Terms of use


Views and Downloads

Views and downloads will return soon






If you are the owner of this record, you can report an update to it here: Report update to this record

TO TOP