User suggests that fast CI/CD pipelines, especially when amplified by AI, need explicit stop conditions and reversibility to prevent deploying mistakes faster. This implies a need for features in CI/CD or deployment tools to manage rapid, potentially AI-driven, deployments more safely.
I wrote a new blog post for the AWS Executive in Residence Blog. "Your AI Coding Assistants Will Overwhelm Your Delivery Pipeline: Here’s How to Prepare" I was inspired by a statement from the latest DORA report: “Agentic AI's primary role in organizations is that of an amplifier. It magnifies the strengths of high-performing organizations and the dysfunctions of struggling ones.” I think this sums up very clearly what I observe in many organizations. They introduce new AI tools in software development and hope that these will work their magic on software development. 🧚♀️ But 1/ coding was never the bottleneck 2/ if you optimize one step in a system, a bottleneck appears somewhere else. And what is unfortunately still the biggest bottleneck when it comes to speed? (For innovation and value, it's something else...) The release and delivery process. Only 22% of companies really do CI/CD. The rest just call it that and release occasionally, rarely, or only when absolutely necessary. And these immature release processes, with manual steps, are now encountering more code, which contains correspondingly more errors. What can be done? 1/ Think systemically. 2/ Observe, measure, and continuously optimize your software delivery process. 3/ Finally do CI/CD properly. In 2026, it's really time to do your homework... What do you think? Leave a comment... ---- Follow me. Learn more about high-performance software development organizations in my book “All Hands on Tech - Your How-To Guide for Building Great Tech Organizations at Scale” https://lnkd.in/digVrbE5