GitLab Launches Duo Agent Platform to Orchestrate AI Across DevSecOps Lifecycle
GitLab Inc., known for delivering one of the most comprehensive and intelligent DevSecOps platforms, has officially announced the general availability of the GitLab Duo Agent Platform. With this launch, GitLab aims to move beyond basic AI-assisted coding and address deeper inefficiencies that slow down modern software delivery.
Over the past few years, AI-powered tools have significantly improved developers’ ability to write code faster. In fact, many teams report productivity gains of up to ten times during the coding phase. However, since developers typically spend only around 20% of their time writing code, the overall impact on innovation speed remains limited. This imbalance has become widely recognized as the “AI paradox” in software development. While coding accelerates, other stages of the lifecycle struggle to keep up.
Latest IT Technology Trends in 2026: How AI Is Transforming Cybersecurity Operations
As a result, many organizations face growing bottlenecks. Faster code creation often leads to mounting backlogs in code reviews, increased security vulnerabilities, additional compliance checks, and more downstream bug fixes. Consequently, teams struggle to convert coding speed into real delivery velocity.
To solve this challenge, GitLab Duo Agent Platform introduces intelligent orchestration and agentic AI automation across the entire software development lifecycle. Importantly, it operates within an organization’s existing context, standards, and governance guardrails, ensuring AI actions remain aligned with internal policies.
With the general availability release, GitLab significantly expands its intelligent AI agent and orchestration capabilities. At the core is Agentic Chat, a context-aware assistant available across the GitLab Web UI and popular IDEs. It uses multi-step reasoning to answer complex questions and autonomously perform actions by drawing insights from issues, merge requests, pipelines, and security findings.
Agentic Chat supports multiple use cases. For analysis, it can create issues, epics, and merge requests while summarizing key insights and offering actionable guidance. From a coding perspective, it generates code, modernizes applications, fixes bugs, creates tests, and produces documentation across multiple languages and frameworks. Additionally, it helps teams understand, configure, and troubleshoot CI/CD pipelines. On the security front, it explains vulnerabilities, prioritizes risks, and recommends fixes to save time and reduce exposure.
Beyond chat-based assistance, GitLab introduces Foundational Agents built by its experts. These include the Planner Agent, which helps teams’ structure and prioritize work, and the Security Analyst Agent, which reviews vulnerabilities and explains their impact in plain language. Moreover, organizations can create Custom Agents through the AI catalog, enabling teams to tailor agents to their engineering standards. GitLab also integrates External Agents, such as Claude Code from Anthropic and Codex CLI from OpenAI, providing native access with transparent security controls.
Latest IT Technology Trends in 2026: How Zero Trust Security Reduces Blast Radius During Active Breaches
In addition, Foundational Agentic Flows automate complex tasks like converting issues into merge requests, modernizing CI/CD pipelines, fixing pipeline failures, and streamlining code reviews. Governance and visibility features ensure leaders can track agent usage, measure impact, and maintain responsible AI adoption.
The platform is available across GitLab.com, GitLab Self-Managed, and GitLab Dedicated as part of the GitLab 18.8 release. Flexible model selection, group-based access controls, and enterprise-grade governance further support scalable adoption. GitLab also introduces GitLab Credits, allowing Premium and Ultimate customers to access Duo Agent Platform features with included monthly credits.
Write to us [wasim.a@demandmediaagency.com] to learn more about our exclusive editorial packages and programmes.
