The 1st International Workshop on Domain-Specific Agentic AI (DSAAI 2026) aims to bring together researchers, industry practitioners, and domain experts to discuss the design, evaluation, deployment, and governance of specialized AI agents and agentic workflows. Agentic AI refers to Artificial Intelligence (AI) systems designed to function as autonomous or semi-autonomous agents. Such systems can understand high-level goals, decompose them into actionable tasks, and execute these tasks with minimal human intervention.

While Generative AI (GenAI) and general-purpose agents have demonstrated impressive capabilities in recent years, deploying these systems in specialized domains such as healthcare, energy, finance, legal practice, software engineering, and industrial robotics reveals significant challenges related to domain accuracy, reliability, compliance, and trustworthiness.

DSAAI 2026 will address key questions including:

• How can generative AI models be adapted for domain-specific applications?
• How should agentic workflows be designed and orchestrated?
• How can agentic systems be integrated into EOSC infrastructures and HPC/HTC environments?
• How can safety, trustworthiness, and regulatory compliance be ensured?
• What are suitable evaluation methodologies for specialized autonomous systems?

Submission Guidelines

We invite researchers, practitioners, and industry experts to submit

• Extended Abstracts (3–4 pages) or
• Short Papers (5–6 pages)

describing original research, practical experiences, innovative methodologies, system designs, and lessons learned from real-world deployments of generative and agentic AI systems.

All papers must be original and not simultaneously submitted to another conference. Submissions must be written in English and follow Springer formatting guidelines.

All submissions will undergo a single-blind peer review process. Evaluation criteria include originality, technical quality, relevance, clarity, practical impact, and the potential to stimulate future research and innovation.

Please proceed to EasyChair via https://easychair.org/conferences/?conf=dsaai2026.

List of Topics

Contributions may include theoretical advances, implementation experiences, industrial applications, AI-enabled knowledge graph solutions, LLM-assisted systems, and novel approaches for building trustworthy AI agents.

• Foundations of Generative and Agentic AI Inference
• Architectural Patterns and Design Considerations for Trustworthy Generative AI
• System Architectures for Serving AI Agents
• Novel Approaches for Unifying Large Language Models and Knowledge Graphs
• Training, Testing and Multi-Agent Orchestration
• Explainability-by-Design Mechanisms for AI Services
• EOSC Interoperability Standards and Services
• Knowledge Graph Applications, Tools and Technologies
• Deployment Strategies for Large-Scale Agentic Systems
• Evaluation Methodologies for Trustworthy Agentic AI Systems
• Biodiversity, Energy, Materials Science, Healthcare and Environmental Use Cases

Committees

Program Committee

• Astrid Nieße, Carl von Ossietzky University Oldenburg, Germany
• Christian Vitale, KIOS Research and Innovation CoE, University of Cyprus
• Dimitar Trajanov, Ss. Cyril and Methodius University in Skopje, North Macedonia
• Dušan Vudragović, Institute of Physics, University of Belgrade, Serbia
• Fernando Ortiz-Rodriguez, Universidad Autónoma de Tamaulipas, Mexico
• Fatmana Senturk, Pamukkale University, Turkey
• Milos Cvetanović, School of Electrical Engineering, University of Belgrade, Serbia
• Milos Jovanovik, Faculty of Informatics, Technical University of Vienna, Austria
• Sarika Jain, National Institute of Technology Kurukshetra, India
• Slavica Boštjančič Rakas, Institute Mihajlo Pupin, University of Belgrade, Serbia
• Petr Knoth, Knowledge Media Institute, The Open University, UK
• Sandra Lovrenčić, Faculty of Organization and Informatics, University of Zagreb, Croatia
• Zaharije Radivojević, School of Electrical Engineering, University of Belgrade, Serbia

Organizing Committee

• Sahar Vahdati – Technische Informationsbibliothek (TIB), Leibniz Universität Hannover, Germany
• Valentina Janev – Institute Mihajlo Pupin, University of Belgrade, Serbia
• Marko Batić – Institute Mihajlo Pupin, University of Belgrade, Serbia
• Lazar Berbakov – Institute Mihajlo Pupin, University of Belgrade, Serbia

Publication

Accepted contributions will be published online through ZENODO before the workshop.

Venue

The conference will be held in the Research and Development Campus of the Institute Mihajlo Pupin, Volgina 15, Belgrade, Serbia.

Contact

All questions about submissions should be emailed to valentina.janev@pipin.rs.

Sponsors

This workshop is supported by: