1st Workshop on Software Engineering in the Age of Agentic AI

The 1st 1-Week Workshop on Software Engineering in the Age of Agentic AI at the Bellairs Research Institute of McGill University in Holetown, Barbados, is taking place from Friday, April 23rd, 2027 to Friday, April 30th, 2027 with a special emphasis on the rapid changes to modern software engineering due to agentic AI.

Fueled by the rise of agentic AI and the accompanying significant reduction of feedback cycle lengths, a transformative shift is occurring in software engineering. This shift introduces both unprecedented productivity gains and fundamental questions about reliability, accountability, and the nature of software development itself. The days where software engineers perform many tasks by themselves seem to be gone. What will be the role of future software engineers? How will future software engineers of all levels of SE expertise interact with their development environment? Will there be software engineers?

Engineering a software system without the help of agentic AI is rapidly becoming the exception rather than the norm. In this context, we are considering the following key discussion topics for our workshop:

  1. Seamless collaboration between agentic AI and human actors – How should tasks be distributed in terms of elicitation, coding, testing, debugging, deployment, and monitoring to create a more efficient and reliable engineering process?
  2. Continuous software engineering as a key enabler for this new world – How can significantly shorter feedback loops support decision-making at any level of software development?
  3. Embracing continuous evolution – How can we shift our mindset from seeing continuous updates by agentic AI as a challenge to leveraging them for faster innovation and better experience for software engineers?
  4. Reliability & transparency – How can we ensure correctness and robustness of software created by agentic AI and how do we understand their decisions?
  5. Security and safety – Does agentic AI apply to safety-critical software or where security concerns are paramount?
  6. Technical debt and long-term sustainability – How can we manage technical debt while maintaining rapid software evolution?
  7. Evaluation & benchmarks – How do we measure the performance, quality, and trustworthiness of agentic AI?
  8. Ethics & responsibility – Who is accountable for AI-generated code and decisions in production systems?

By bringing together experts from various domains, we aim to strengthen the research community, identify best practices, and explore new research directions. Possible outcomes include: a conceptual model for the collaboration of agentic AI and human actors, a comprehensive vision on the topic, case studies, …! The workshop’s ultimate goal is to embrace the reality of agentic AI, transforming it from a risk into a strategic advantage for software engineering.