Harness Extends CI/CD to AI Agents, Bridging Traditional DevOps with AI Development

Harness Inc., a prominent integrated software delivery platform provider, has announced the general availability of its Agent DLC, a new offering designed to bring continuous integration and continuous delivery (CI/CD) practices to the development and deployment of artificial intelligence agents. This new suite provides purpose-built evaluation, deployment governance, and security controls specifically tailored for the unique characteristics of AI agents. The core innovation lies in allowing organizations to manage their AI agents using the same robust CI/CD pipelines they currently employ for traditional software code. This development is highly significant for DevOps and AI practitioners. The inherent non-deterministic nature of AI agents has historically posed considerable challenges for traditional software development lifecycle (SDLC) processes, particularly in testing and deployment. While conventional software yields predictable outcomes, AI agents make autonomous decisions, requiring different validation and operational strategies. Harness's Agent DLC directly addresses this gap, enabling the secure and reliable operationalization of AI agents. This is critical for accelerating the adoption of AI agents in production environments, especially given that internal data suggests only 8% of organizations have successfully deployed agentic AI into production despite widespread interest. The launch of Agent DLC by Harness fits squarely within the broader industry trend of converging AI development with established software engineering methodologies, often referred to as MLOps or AIOps. As AI models and agents move beyond experimental phases and become integral components of enterprise applications, the demand for robust, scalable, and secure operational practices becomes paramount. This move reflects a growing recognition that AI development cannot remain siloed but must integrate seamlessly into existing enterprise-grade workflows to ensure reliability, auditability, and compliance. Other platforms and tools are also evolving to support AI-driven DevOps, highlighting the industry's collective effort to mature the AI development lifecycle. For practitioners, this means a tangible path to applying familiar CI/CD best practices—such as version control, automated testing, and controlled deployments—to their AI agent initiatives. This will lead to improved auditability, faster iteration cycles, and a reduced risk profile for AI-driven applications. Teams should evaluate how their current CI/CD infrastructure can integrate with or adapt to solutions like Agent DLC. It also underscores the necessity for developing new testing paradigms that can effectively validate the non-deterministic behaviors of AI agents, moving beyond traditional unit and integration tests to incorporate more sophisticated evaluation metrics and simulation environments.