Workshop on Agentic AI Systems Development

Held in conjunction with the 31st ACM International Conference on Intelligent User Interfaces (IUI 2026)

March 23–26, 2026 — Paphos, Cyprus

Announcements

October 8, 2025. We are excited to host the first AgentCraft workshop at IUI 2026! Please check out our Call for Participation and submit your work by December 19!

Overview

Ambitious efforts are underway to build AI agents powered by large language models (LLMs) across domains including programming, productivity, creative work, and enterprise automation. Despite recent progress, major challenges remain in ensuring agent autonomy, interpretability, oversight, and real-world reliability. Developers struggle to understand and debug agent behavior, to decide when human input is needed, and to evaluate usefulness beyond narrow benchmarks.

The AgentCraft workshop will bring together researchers, developers, and designers to address these challenges. Our focus is on equipping developers with the tools, methods, and evaluation frameworks needed to design, test, and deploy agentic systems responsibly.

Workshop Goals & Format

Our primary goal is to identify a set of guidelines and best practices for developing and debugging LLM-based agentic AI systems. Central to this challenge is understanding how to develop and evaluate agentic AI systems in a human-centered way that includes actual users rather than solely evaluating agent performance on artifical benchmarks that purport to simulate real-world tasks.

The workshop will be a full-day mini-conference including:

We aim to build a community of researchers, developers, and designers spanning academic and industry research to encourage new ideas and new collaborations to develop.

Submissions

Submissions will take the form of research or demo papers, following CEUR publication guidelines. Please see the Call for Participation for relevant topics and submission instructions.

Key Dates