Enterprise Architecture Has Four Domains. The Agentic Enterprise Needs a Fifth.
Why TOGAF, Zachman, and SABSA can’t model an agent — and the domain that can.
Somewhere in your enterprise right now, an AI agent is acting with delegated authority. It is calling APIs, moving data, and composing tools to finish a task no one scripted end-to-end.
Four Domains, Thirty Years
And when your architecture review board asks which domain it belongs to, there is no honest answer.
Enterprise architecture has four domains. Business. Data. Application. Technology. Security runs across all of them. For thirty years, that model held, because everything we built was a system the enterprise owned, configured, and ran.
An agent is not a system you run. It is an actor you delegate to.
The Frameworks Describe What You Build
That distinction breaks the frameworks. TOGAF, Zachman, SABSA, FEAF — every one of them describes what the enterprise builds. None describes a non-human actor that holds delegated authority, acts autonomously, and reaches for tools it was never explicitly granted.
We have spent three decades modeling applications. We have no domain for a workforce.
A Workforce With No Org Chart
The enterprise has hired a workforce it never put on an org chart — and no framework tells it where that workforce lives.
The Fifth Domain
That gap has a name: Enterprise Agent Architecture. A fifth domain, alongside the other four, for the agent workforce itself. It has four layers:
The workforce — which agents exist, what authority they hold, how it is delegated and revoked.
Capability — what they are permitted to touch, enforced rather than assumed.
The control plane — where policy meets runtime: the moment an action is allowed, denied, or held.
Governance — who owns the agents, who audits them, and who is accountable for decisions a human did not make.
Not a Security Problem, Not an Integration Problem
This is not a security problem you can hand to the CISO, or an integration problem you can hand to the platform team. It crosses both, which is precisely why it is an architecture problem.
An Attempt to Be Early
The frameworks will catch up. They always do — late, after the enterprise has already absorbed the shift in pieces and paid for the gaps. This is an attempt to be early.
Enterprise architecture has four domains. The agentic enterprise needs a fifth. Over the coming pieces, I’ll draw it, domain by domain — starting with why TOGAF cannot model a workforce that isn’t human.
The Bridge to Part 1
The fifth domain is only a claim until each of its layers is drawn. Part 1 takes the first one: the Agent / Workforce layer — and why TOGAF cannot model a workforce that isn’t human.
About This Series
Michael K. Saleme — Enterprise Agent Architect
Part 0 is the position paper of the series: the argument that agents constitute a fifth domain of enterprise architecture rather than a new application tier. Every later part develops one layer of that domain.
This page is the canonical version of an essay also published to the Enterprise Agent Architecture newsletter.
Cite the Position Paper
This essay develops the Enterprise Agent Architecture position paper, which is published and peer-citable on Zenodo under CC BY 4.0. Cite the concept DOI — it always resolves to the latest version.
Saleme, M. K. (2026). Enterprise Agent Architecture: The Case for a Fifth Architecture Domain for the Agentic Enterprise. Zenodo. https://doi.org/10.5281/zenodo.21105314
BibTeX
@misc{saleme2026eaa,
author = {Saleme, Michael K.},
title = {Enterprise Agent Architecture: The Case for a Fifth Architecture Domain for the Agentic Enterprise},
year = {2026},
publisher = {Zenodo},
doi = {10.5281/zenodo.21105314},
url = {https://doi.org/10.5281/zenodo.21105314}
}
Michael K. Saleme
Enterprise Agent Architect · Cognitive Thought Engine