Workshop at CASCON 2026 · Toronto
The Future of Software-Intensive Systems
About the workshop
Agentic AI marks a paradigm shift in the engineering of software-intensive systems. Unlike conventional, deterministic applications, agentic systems are decentralized software ecosystems in which specialized autonomous agents use Foundation Models — LLMs and SLMs — as cognitive reasoning engines, planning, deciding and learning to pursue shared goals through coordinated autonomy. This workshop explores the symbiotic relationship between Software Engineering and Agentic AI, in both directions.
Direction 01
Autonomous agents are transforming the software engineering lifecycle — accelerating development and automating complex operations, while making software-intensive systems significantly harder to validate, certify, maintain and evolve.
Direction 02
The operational uncertainty inherent to agentic ecosystems demands new design abstractions, reference architectures and runtime engineering techniques to build, operate, govern and evolve resilient multi-agent systems.
Why agentic systems are hard to engineer
Agents rely on probabilistic reasoning models, so their behaviour is non-deterministic by design.
They depend on highly uncertain context: a volatile environment and the unpredictable, emergent behaviour of coordinating agents.
Yet they must reliably remember past actions, track multi-step progress toward shared goals, and keep context across agents, tools and users.
Goals
The workshop brings together researchers and practitioners around the engineering of software-intensive systems in the Agentic AI era.
Discuss the challenges and opportunities of engineering software-intensive systems with Agentic AI, across the full lifecycle.
Examine critical runtime challenges: dynamic context management — including memory staleness and context relevance — adaptive control, self-adaptation, runtime governance and viability-zone enforcement under high operational uncertainty.
Connect researchers, industrial practitioners and students to foster cross-sector partnerships, joint research initiatives and collaborative networks at the intersection of Agentic AI and Software Engineering.
Topics
In software engineering, self-adaptive systems, autonomous agents, context-aware computing, runtime engineering and related fields.
Ready to share experiences and challenges from engineering software-intensive systems in the Agentic AI era.
Looking to anchor their research in a fast-moving, high-impact domain.
Program
· · , . The workshop moves from keynote-style talks, to discussion in working groups, to a plenary synthesis of what the groups found.
The workshop is part of the CASCON 2026 workshops program.
Speakers
Invited talks open the workshop with perspectives that frame the challenges and set the agenda for the working groups. Each talk's abstract and the speaker's bio are in the program.
We are inviting leading researchers and practitioners to share their perspectives on engineering with — and of — Agentic AI. Confirmed speakers will appear here as they are announced.
Contact the organizers →Organizers
The workshop is organized by researchers and practitioners from the University of Victoria, York University and IBM.
Attend
The workshop is part of CASCON 2026, the , held at in Toronto. Attendance is through CASCON registration.
Questions about the workshop? Write to the organizers.