
November 23โ27, 2026
๐๐น๐ผ๐ฏ๐ฎ๐น ๐๐ ๐๐ฒ๐น๐ฒ๐ด๐ฎ๐๐ฒ ๐๐ผ ๐๐ด๐๐ฝ๐ - ๐๐ ๐๐ฑ๐๐ฐ๐ฎ๐๐ถ๐ผ๐ป & ๐ช๐ผ๐ฟ๐ธ๐ณ๐ผ๐ฟ๐ฐ๐ฒ ๐๐ฒ๐๐ฒ๐น๐ผ๐ฝ๐บ๐ฒ๐ป๐ ๐๐ผ๐๐ป๐ฐ๐ถ๐น ๐ ๐ฒ๐บ๐ฏ๐ฒ๐ฟ
Global Alliance for Artificial Intelligence
Agentic AI is moving beyond task automation toward systems that can make recommendations, initiate actions, and increasingly operate with delegated authority. Yet most organizations still govern AI primarily at the model or system level, leaving a critical question unanswered: who gets to give an AI agent authority, and under what conditions? This session introduces a practical approach to delegation governance, examining how organizations can define decision boundaries before granting agents the ability to act. It explores five essential mechanisms: delegated authority, decision boundaries, intervention rights, escalation, and authority reclamation. It also considers how accountability should be maintained when an AI agent operates across workflows involving multiple teams and stakeholders. The session argues that responsible agentic AI requires more than accurate models or human-in-the-loop controls. Organizations need explicit rules governing what an agent may decide, when humans must intervene, who remains accountable, and how authority can be withdrawn. The goal is to provide practitioners with a practical framework for thinking about agentic AI as an organizational delegation problem, not simply a technology deployment problem.
See every session across the AI Agents Conference November 2026 line-up.
View Agenda