GitLab 18.5 accelerates Agentic AI: from experimentation to wider deployment.

Agentic AI in GitLab: a deliberate evolution since version 18.

Agentic AI is not a sudden innovation. Since GitLab 18.0, GitLab has been developing AI agents step by step that:

• Understand context about repositories, pipelines and policies

• Support developers with targeted actions

• Able to perform tasks increasingly autonomously, within secure frameworks

Earlier releases mainly emphasised individual productivity and experimentation. GitLab 18.5 brings these developments together, allowing AI agents to now be used more broadly in CI/CD, security and governance, while maintaining control and auditability.

Why GitLab 18.5 is a key moment.

Where previous 18.x releases offered experimental features, 18.5 provides consolidation and expansion:

• From support to broader deployment:AI agents can now perform multiple steps in workflows, such as prepare merge requests or support security tasks.

• Platform-wide integration:Agentic AI is integrated into pipelines, policies and governance, no longer limited to IDE or chat.

• Enterprise-ready and manageable: Better insight into AI's actions, audit trails and policy-driven usage make agents suitable for larger organisations.

• Enterprise-ready and manageable: Better insight into AI's actions, audit trails and policy-driven usage make agents suitable for larger organisations.

Note that the agents are still largely beta, and human supervision remains essential.

What does this mean concretely for your organisation?

For tech-mature organisations, GitLab 18.5 translates into tangible benefits:

• Shorter lead times without additional pressure on teams

• Less manual work in reviews, security and pipelines

• Consistent adoption of standards and best practices

• More room for architecture, product development and innovation

• More room for architecture, product development and innovation

Agentic AI absorbs complexity so that teams can focus on value creation.

For whom is this release relevant?

GitLab 18.5 is of particular interest to organisations that:

• working with multiple DevOps teams and repositories

• high demands on security and compliance

• want to scale up without linearly hiring more people

• using GitLab as a strategic platform, not just as a tool

GitLab 18.5 is accelerating the evolution of Agentic AI.

Agentic AI began with GitLab 18 as its vision and foundation. GitLab 18.5 accelerates this functionality and makes it more widely applicable, with new agent types and integrations. Agents remain largely in beta/experiment status, so human control remains crucial. For organisations looking to move forward with smart AI support in DevSecOps, now is the time to get acquainted and experiment with Agentic AI.