Call for Papers
Advancing agentic AI workflows from concepts to dependable systems
AAWT 2026 invites original submissions on the methods, architectures, platforms, tools, and practices needed to build and evaluate agentic AI workflows and technologies. The workshop welcomes research and practice on systems where AI agents plan, reason, call tools, coordinate tasks, interact with humans, access external services, and execute complex workflows across software, data, business, scientific, and industrial environments.
Agentic AI is rapidly changing the way digital work is organized and automated. New AI agents can support information seeking, software development, decision support, data analysis, customer service, operations, research, and many other tasks. However, building agentic systems that are reliable, transparent, secure, controllable, and useful in real settings remains a major challenge.
AAWT focuses on the full lifecycle of agentic workflows, from design and modelling to orchestration, monitoring, evaluation, deployment, governance, and maintenance. The workshop is interested in both single-agent and multi-agent systems, human-in-the-loop workflows, agent communication protocols, tool integration, memory and context management, workflow verification, and operational risks.
The workshop welcomes full research papers, short papers, position papers, experience reports, industrial case studies, tool demonstrations, benchmark papers, datasets, and papers describing lessons learned from real-world agentic AI deployments.