When GitLab saw an opportunity to reimagine software engineering as a human and AI collaboration, its engineering and product teams came together to do just that.
They created GitLab Duo Agent Platform, a DevSecOps orchestration platform, launched in public beta in July 2025. Designed to support developers at every stage of the software development lifecycle, the solution offers numerous features, including the GitLab Knowledge Graph, which transforms how developers and agents understand and navigate complex codebases.
Principal Product Manager Jordan Janes, who helped develop GitLab Duo Agent Platform, believes that the solution reflects the key pillars of GitLab’s product development process: collaboration, iteration, and a deep focus on responsible AI. He and his teammates see these values in action when building new solutions, relying on each other’s expertise — and trust — to help ensure every product meets customers’ needs.
According to Janes, working on GitLab Duo Agent Platform has been both fast-paced and rewarding, reflecting the ambitious goals that each team sets for itself and the wider organization. As he and his peers continue to find new ways to enhance the platform, they’re focused on defining priorities while leaning on each other along the way.
“We view ourselves as one collective team working toward near-term and long-term goals,” Janes said.
Read on to learn more about the launch of GitLab Duo Agent Platform, the impact it has on GitLab’s customers and how it reflects the heart of the company’s culture.
Provide an overview of the launch of GitLab Duo Agent Platform — who was involved, when it took place, and why it was significant for GitLab?
GitLab Duo Agent Platform is the priority for the AI teams here at GitLab. We’ve been focused on improving the platform each week and with each milestone, and we’re hearing exciting feedback from early customers. A few key advancements include the AI Catalog, where teams create and share custom agents, along with the GitLab Knowledge Graph, which provides intelligence to help developers and agents navigate complex codebases.
We also introduced Agent Sessions, which provide detailed logs and metadata for transparency into how agents are operating across GitLab. With these advancements, customers can further accelerate tasks across the software development lifecycle. All of these improvements were supported by multiple teams working together.
What was your role in this milestone? How did you and your team contribute to its progress, and what perspective did you gain from being part of the process?
Collaboration is one of GitLab’s core values, and product managers have the opportunity to work cross-functionally with teams across the entire company. For me, this has always been one of the best parts of the role. My product management teammates and I have a strong influence on defining the goals for each milestone, the overall customer experience, and helping orchestrate the internal teams to deliver. Particularly in the AI space, we apply our Iteration core value to make consistent progress while ensuring we’re moving toward a longer-term vision. We also meet directly with customers, hear their feedback, and use it to inform our planning.
What challenge or opportunity did this milestone address? How has reaching it influenced GitLab’s approach to innovation or collaboration moving forward?
This milestone has helped mature GitLab Duo Agent Platform as we work toward general availability readiness. Features such as our AI Catalog, improved Agentic Chat, and the Knowledge Graph help enhance our product’s capabilities. Each of these features is supported by multiple teams that collaborate daily.
These product areas are the result of early bets, where we experimented internally to evaluate technology choices and the target customer experience. GitLab has an experimentation and innovation culture, and it’s rewarding to see that culture in action in our AI capabilities.
“GitLab has an experimentation and innovation culture, and it’s rewarding to see that culture in action in our AI capabilities.”
Share memorable moments from along the way — and explain why each was meaningful.
Each milestone’s release date is memorable. Our go-to-market teams do an excellent job of highlighting improvements and communicating why they benefit customers. Across demos, blog posts, and videos, I’m always excited to see the presentation of what we’ve been working on come to life.
I also had a memorable moment of teamwork where a teammate stepped in to cover an internal presentation I was scheduled to give so I could attend a last-minute customer meeting. It’s a small moment, but feeling confident that I can depend on my teammates is invaluable.
My final memorable moment was introducing the GitLab Knowledge Graph. The engineering team has done an incredible job building a rich code intelligence system that is easy for customers to use.
How did the milestone conclude? Describe the atmosphere across the team and any reflections you have on what this experience reveals about GitLab’s culture and approach to building AI-driven products.
Each milestone concludes by us proudly releasing our changes, then collectively moving into execution for the next milestone! Working on GitLab Duo Agent Platform is fast-paced, and it’s rewarding to see a product mature from an early experimental version into beta with broader customer adoption and toward general availability. We are constantly considering improvements and functionality that we can deliver now and what to prioritize next. This is where we apply our Results for Customers core value, ensuring that all our efforts are focused on helping our customers succeed.
How did collaboration across teams contribute to the success of this milestone, and what does it show about GitLab’s culture of transparency and iteration?
We set ambitious goals for ourselves and across the organization. We’re transparent about progress across our initiatives, which gives us the space to flag risks and ask for help across teams. We view ourselves as one collective team working toward near-term and long-term goals. This is a different mindset than individual subteams with strict areas of ownership. This really helps us collaborate at scale and iteratively deliver improvements.
Frequently Asked Questions
What is the GitLab Duo Agent Platform, and why is it significant?
Launched in public beta in July 2025, the GitLab Duo Agent Platform supports developers throughout the software lifecycle. It includes innovations like the AI Catalog for creating and sharing agents, Agent Sessions for transparency, and the GitLab Knowledge Graph for navigating complex codebases.
How did collaboration shape this milestone?
The project was built through close cross-functional teamwork across engineering, product and go-to-market teams. Teams shared progress openly, supported each other when challenges arose and applied GitLab’s iteration value to deliver consistent, customer-focused improvements.
What makes GitLab’s product development process unique?
GitLab’s culture emphasizes experimentation, responsible AI, and iteration. Teams are empowered to test ideas, make early bets, and refine them quickly based on customer feedback — fostering innovation while maintaining transparency.
How does the platform reflect GitLab’s culture?
From internal collaboration to customer-focused milestones, the platform’s creation embodies GitLab’s core values: Collaboration, Iteration, and Results. Team members trust one another, share ownership, and view themselves as one collective unit working toward shared goals.
What was a memorable moment from the development process?
For Principal Product Manager Jordan Janes, milestones like launching the GitLab Knowledge Graph and seeing cross-team support in action stand out as testaments to teamwork and trust — two pillars of GitLab’s culture.
