ACTION CENTRIC AND HYBRID AGENTIC AI IN THE GOVERNMENT AND COMMERCIAL DATA STACK
Agentic AI does not operate in isolation. It is embedded within enterprise data stacks that define what agents can observe, what actions they are allowed to take, and how success is measured.
In Part II of this series, Action Centric and Hybrid Agentic AI in the Government and Commercial Data Stack, Ricci Mulligan moves from conceptual clarity to operational reality. The paper examines how action-centric and hybrid agentic systems function within governed cloud environments, data warehouses, data lakes, graph models, and language-mediated interfaces across government and commercial settings.
Rather than focusing on specific tools or vendors, this paper explores how architectural choices shape rationality, authority, and governance. It clarifies why action-centric systems excel at disciplined execution, why hybrid architectures emerge to surface epistemic limits, and how language models influence interpretation without redefining decision authority.
Author: Ricci Mulligan, Alyned, Former Acting Principal Deputy Assistant Secretary, OIT, Department of Veterans Affairs (Retired)