AGENT 2027

Agentic engineering focuses on the design, development, evaluation, and operation of agentic AI systems that exhibit goal-directed autonomy, reasoning, tool use, interaction, and continuous evolution. Foundation-model-based agents are moving from prototypes into real software engineering workflows, including development, testing, maintenance, operations, and knowledge-work platforms.

These systems can decompose goals, use tools, call APIs, coordinate with humans and other agents, generate or modify software artifacts, monitor execution, and operate in partially open-ended environments. However, many current agentic AI applications are assembled from prompts, tools, memory stores, orchestration frameworks, model-context protocols, and ad hoc evaluations without systematic treatment of requirements, architecture, verification, or oversight.

AGENT’27 will provide an interactive forum for researchers and practitioners to advance engineering foundations for agentic AI systems. The workshop will emphasise empirical evidence, reusable methods, operational lessons, and open research problems across the agentic AI system lifecycle.

The anticipated goals and outcomes are to:

• Refine a research agenda for agentic engineering, including open problems where current software engineering theories, methods, and tools are insufficient.

• Identify validation objectives, empirical study designs, datasets, benchmark needs, and artifact-sharing opportunities.

• Share practical lessons from deployed or near-deployed agentic AI systems, including reliability failures, safety controls, human oversight, observability, and AgentOps practices.

• Foster cross-community collaboration across requirements, architecture, testing, responsible AI, and industrial AI-enabled software development.

Call for Papers

Agentic engineering focuses on the design, development and operation of agentic AI systems that exhibit goal-directed autonomy, reasoning, tool use, and continuous evolution. AGENT’27 invites researchers and practitioners to submit original research, experience, position, tool, benchmark, and vision papers on software engineering foundations for agentic AI systems.

Topics of interest include, but are not limited to:

• Requirements engineering, goal modeling, and specification for agentic AI systems;

• Architecture, orchestration, tool use, memory, protocols, and multi-agent coordination;

• Verification, validation, testing, red-teaming, simulation, and benchmarking;

• AgentOps, observability, runtime monitoring, intervention, and incident response;

• Responsible AI, safety, security, privacy, accountability, and governance;

• Agentic systems for requirements, design, coding, testing, deployment, maintenance, and operations;

• Human-agent interaction, collaboration, oversight, trust calibration, and socio-technical effects;

• Empirical studies, industrial case studies, datasets, tools, and deployment lessons.

Submission Types

We welcome:

• Full papers (research or experience): up to 8 pages, excluding references.
• Short papers (position, emerging research, tool, benchmark, or experience): up to 5 pages, excluding references.

All submissions must be original, written in English, and formatted according to the IEEE conference proceedings template. Submissions will be reviewed single-anonymously. Accepted workshop papers will be published by IEEE.

The official publication date of the workshop proceedings is the date the proceedings are made available by IEEE. This date may be up to two weeks prior to the first day of ICSE 2027. The official publication date affects the deadline for any patent filings related to published work.

Important Dates

• Workshop paper submission deadline: November 27, 2026

• Notification of acceptance: December 11, 2026

• Camera-ready deadline: January 29, 2027

All dates are Anywhere on Earth (AoE).