How AI-Created Code Will Strain DevOps Workflows
Mike Vizard shares insights from Nick Durkin on how the increasing reliance on AI-generated code impacts DevOps workflows, highlighting necessary pipeline changes and the importance of modern engineering governance.
How AI-Created Code Will Strain DevOps Workflows
By Mike Vizard
Nick Durkin, Field CTO at Harness, offers a perspective on how the growing use of artificial intelligence (AI) for code generation is introducing new challenges to DevOps workflows. As organizations turn to AI tools to assist in software development, existing pipeline bottlenecks risk being exacerbated, leading to a need for re-engineering the way DevOps is conducted.
Growing Bottlenecks from AI Code Generation
Durkin notes that with AI generating larger volumes of code, the speed of delivery isn’t necessarily increasing. Industry reports show slowdowns and a rise in bugs—even as code production ramps up. Addressing these challenges means simplifying workflows for engineers and implementing automated guardrails to prevent errors, whether code originates from humans or AI.
The Benefits of Template-Driven Pipelines
One major issue is the proliferation of distinct pipelines for each application, resulting in thousands of difficult-to-manage workflows. Durkin advocates for reusable, template-driven pipelines incorporating policies that ensure failures are flagged early and can be remedied quickly. He compares this approach to a video game that encourages iterative improvement, as opposed to rigid, static approval processes.
The Need for Modernized CI/CD Platforms
The rise of AI accelerates the push to update CI/CD platforms. Durkin recommends moving away from stitching together disparate tools towards scalable, automated, and reusable pipelines. Here, AI can assist in automated testing, troubleshooting, and infrastructure management—reducing repetitive tasks so engineers focus on higher-value work.
Platform Engineering and Organizational Change
Despite AI’s ability to speed up certain tasks, Durkin points out that overall delivery timelines may stretch due to increasing pipeline complexities. He aligns this with a broader trend towards platform engineering—where context, policy, and governance take precedence over raw speed. Conversations about software delivery are extending beyond engineering to include CEOs and executive boards.
AI as an Accelerator for Disciplined Engineering
Ultimately, Durkin stresses that AI will not replace the fundamentals of robust software delivery; rather, it will change how these fundamentals are achieved. The companies that succeed will treat AI as a tool to accelerate disciplined engineering, not as a shortcut. Effective pipelines, policies, and platform practices will be key in ensuring the reliability and security of applications amid the evolving technological landscape.
For more insights and video interviews, visit DevOps.com.
This post appeared first on “DevOps Blog”. Read the entire article here