AI and Quality – A Contradiction? Micro-Agents as a Path to Sustainable Software Engineering

The omnipresence of AI solutions is undeniable, raising hopes of better solutions with less effort. However, AI-generated code often results in 'workslop' and degraded quality. Studies have shown an increase in duplication and technical debt. Furthermore, the MISRA standard now requires that AI code meets the same compliance requirements as human-written code, including all the associated obligations for testing, traceability and quality assurance. This talk will explore how micro-agent architectures for tasks such as test generation or code quality monitoring offer a sustainable alternative to monolithic AI use. This is because they enable the controlled, context-aware integration of AI that supports software quality instead of compromising it.

Value for the audience:
How to apply AI meaningfully without compromising quality.

Problems addressed:
Degradation of Code Bases

Useful integration of AI into Software Development

Talk language: English
Level: Advanced
Target group: Software Developers, Quality Managers, Test Developers, Test Managers

Company:
Vector Informatik GmbH

Dipl.-Inform. Andreas Horn