Keynotes
Empirical Research for AI-Native Software Engineering
Markku Oivo
Abstract
Software engineering is being revolutionized by generative AI faster than our evidence can mature. The unit of analysis in empirical research is shifting from individual developers to rapidly evolving human-AI systems, whose behavior is stochastic, opaque, and version-dependent. For a community built on careful measurement and reproducible claims, this is both an opportunity and a quiet crisis. Which findings about AI-native development are durable knowledge, which are perishable snapshots. AI-native software engineering challenges construct validity, as productivity, quality, and expertise change meaning; external validity, as results depend on rapidly obsolete tools; and reproducibility, as the systems studied are proprietary, non-deterministic, and often impossible to replicate.
U(b)locking AI-native Development for Enterprise Software
Tobias Schimmer
Abstract
SAP’s AI-native developer experience program extends across our full product development lifecycle — requirements, architectural and user experience design, coding, testing, operations, and support — serving all people in development roles, including engineers, product managers, and customer support. In this keynote, we share insights on how we are actively experimenting and evolving our AI-native PDLC practices while simultaneously removing major blockers and frictions for AI and human effectiveness leveraging infrastructure, frameworks, and automation; all of which is backed by data-driven insights from surveys, telemetry, and communities.