GitLab used its Transcend virtual event to outline an “intelligent orchestration” approach to using AI agents across the software development lifecycle. It also announced a new assessment programme and a community hackathon for building custom agents.

GitLab positioned agentic AI as a way to automate routine tasks beyond coding, spanning planning, development, testing, security, and deployment. It also emphasised governance controls for organisations adopting AI in software delivery.

Product Direction

During the event, Chief Executive Officer Bill Staples described what he called the “AI paradox” in software delivery. He said AI tools can significantly increase coding output, but coding makes up only a small part of a developer’s day, limiting overall gains in delivery speed.

GitLab’s response is to extend automation across the entire lifecycle rather than focusing on the code editor. Its orchestration model aims to align AI agent actions with organisational workflows, compliance requirements, and standards.

Chief Product and Marketing Officer Manav Khurana outlined three elements of the strategy: an “Agentic Core”, a combined DevOps and security workflow, and enterprise guardrails. GitLab described the Agentic Core as a new layer that brings together its GitLab Duo Agent Platform with a single context model for AI agents.

Khurana said piecemeal adoption of AI features would not be sustainable for organisations with broad lifecycle needs.

“The reality is teams want AI for hundreds of use cases across the software lifecycle, and adding AI feature by feature simply doesn’t scale,” Khurana said. “With GitLab’s platform approach, teams can orchestrate AI agents across planning, development, testing, security, and deployment using the same context, permissions, and security model. That’s how I see AI becoming operational, shareable, and governed across your organisation.”

Customer Examples

Sessions included a customer spotlight with Southwest Airlines, which discussed using the GitLab Duo Agent Platform in software delivery. GitLab linked the use case to airline operational requirements, including reliability expectations tied to round-the-clock operations.

In a separate session, GitLab Vice President of Customer Experience Sherrod Patching discussed results from organisations including Ericsson, Deutsche Telekom, and Barclays. GitLab said the examples showed how organisations are approaching software delivery modernisation and measurement.

GitLab also announced an assessment programme it expects to launch next month. The programme focuses on measuring software delivery maturity and mapping a modernisation path. GitLab described it as a structured way for organisations to understand current practices and introduce more automation and controls over time.

Partner Activity

GitLab highlighted its relationship with Oracle Cloud Infrastructure, describing how the companies work together on deployments on Oracle’s cloud. The event also referenced a managed services programme with Data Intensity as part of that delivery model.

Oracle’s Victor Restrepo said customers are looking for managed operations alongside security and service guarantees.

“Our managed services program with Data Intensity allows customers to focus on innovation instead of operations by delivering a fully managed GitLab experience on OCI with enterprise-grade security and SLAs,” Restrepo said. “Together, GitLab’s intelligent orchestration and OCI’s cloud economics give organisations the flexibility and performance they need to succeed.”

GitLab’s partnership messaging emphasised pre-validated deployment patterns and the economics of running workloads on Oracle’s infrastructure. The event also positioned the partnership as an option for organisations seeking specific deployment and control choices when adopting AI-driven software delivery workflows.

Agents And Metrics

Product demonstrations showed AI agents operating across multiple stages of software delivery, rather than within a single tool. GitLab said the demos also provided visibility into AI’s impact on software engineering metrics.

This focus reflects a broader shift in how technology leaders assess AI in development environments. Organisations increasingly want evidence that AI-driven changes affect throughput, quality, and risk, and clarity on how automation fits into existing controls, including security scanning and compliance checks.

Developer Hackathon

Alongside product updates, GitLab launched a virtual hackathon for developers to build custom agents and workflow “flows” for reuse. GitLab said entries will be shared with the community, with selected projects receiving a permanent listing in GitLab’s AI Catalog.

The hackathon is part of a broader push to grow an ecosystem around reusable agent patterns. If widely adopted, these patterns could shape how teams standardise AI use across projects, including how they apply permissions and consistent context across development and security tasks.

GitLab said the assessment programme and hackathon support its broader push for agentic AI orchestration across the software lifecycle, with further updates expected as customers expand AI usage beyond coding into testing, security, and release management.