Enterprise IT leaders are warming to DevSecOps AI agents from GitLab and Harness, drawn by their deep integration with platform data that provides critical context for autonomous tasks. GitLab’s Duo Agent Platform reached general availability on January 15, 2026, for Premium and Ultimate subscribers on GitLab.com and self-managed deployments, with Dedicated SaaS to follow. The platform equips agents with comprehensive project context—including source code, merge requests, epics, users, and access rights—to enable informed decisions aligned with organizational standards, as stated in GitLab’s announcement reported by GitLab Blog.

Manav Khurana, Chief Product and Marketing Officer at GitLab, emphasized, “Our customers experienced GitLab Duo Agent Platform during its beta period, and we were able to satisfy a variety of quality checks across different deployment options before its general availability. Enterprise customers who choose self-hosted model deployments for data sovereignty or compliance requirements will see comparable AI performance as we continue to enhance our offering.” This rollout includes foundational agents like the planner and security analyst, alongside Agentic Chat for workflow analysis, code generation, testing, CI/CD pipeline management, and vulnerability remediation.

Harness, meanwhile, expanded its AI SRE module with the Human-Aware Change Agent, which weaves human conversations from chat and video into machine data like tickets and software changes during incidents. An automotive company using Harness AI SRE in beta slashed incident resolution from over 60 minutes to 2-3 minutes by integrating CI/CD and telemetry data, according to TechTarget.

Context as the Enterprise Hook

Anuj Tyagi, Senior Site Reliability Engineer at an unnamed communications company, tested GitLab Duo and noted, “It not only generates code, which a lot of other IDEs or IDE plugins can do, but it also tightly integrates with a code repository to support fixing pipeline failures and easy-to-understand security reports. It’s much better than I was expecting, especially how it understands context for a large repository and solves pipeline issues.” This platform grounding—GitLab’s unified data model versus Harness’s knowledge graph from SaaS pipelines—sets these agents apart from generic tools.

GitLab’s agents support custom and third-party integrations like Anthropic’s Claude Code and OpenAI’s Codex CLI, connecting via Model Context Protocol (MCP) servers to external systems such as Jira and ServiceNow. Pricing uses GitLab Credits, with Premium users getting $12 monthly per user and Ultimate $24, refreshing automatically to encourage broad adoption without extra costs initially, as detailed by GitLab Investor Relations.

Harness’s platform, generally available since August 26, breaks tasks into single-step components assigned to specialized agents, with evaluation agents ensuring quality. It boasts production use by 100 of its 1,000 enterprise customers for pipeline building, addressing the ‘AI bottleneck’ where AI-generated code surges but delivery stability drops 7.2%, per CEO Jyoti Bansal in TechTarget.

Production Wins and Real-World Gains

Roger Blakely, Fractional CIO at StratITech, highlighted Harness’s modularity: “In some cases, it won’t make sense to consolidate. The beauty of Harness is, you can pick and choose what you want, and the AI SRE can still hook into the others and still get the information required for the incident management process.” An automotive firm echoed this, planning gradual consolidation while valuing flexibility during transition.

GitLab’s GA in GitLab 18.8 includes the Duo Planner Agent and Security Analyst Agent, with auto-dismissal of irrelevant vulnerabilities. Early adopters like Southwest Airlines and Oracle Cloud scale AI governance via context-aware assistance for pipeline fixes and code reviews, as noted in WebProNews. Self-hosted performance has improved, though Tyagi observed initial gaps in prompt understanding compared to SaaS.

Torsten Volk, Analyst at Omdia, praised Harness: “Human-AI collaboration is the most important key to AI success in general. The Harness SRE agent is very interesting in that regard, as it goes far beyond just adding Slack and war rooms as just another MCP source for agents to draw from, but it proactively has agents participate in communication flows.”

Rivals and Strategic Edges

Competition intensifies with GitHub’s Copilot Extensions serving 180 million developers, relying on third-party integrations unlike GitLab’s unified approach. Jim Mercer of IDC observed, “In contrast to GitLab’s focus on a unified platform, GitHub relies on Copilot Extensions to ease integration with third-party tools and extend capabilities across the pipeline.” GitHub’s retracted self-hosted Actions pricing ($0.002/min) spooked some firms, creating openings, per Blakely: “I know of two companies that have backed away from that because they think that [the pricing change] potentially will come back at some time in the future.”

Harness differentiates with opinionated templates embedding security and compliance, contrasting GitLab’s customizable agents. IDC’s Mercer added, “Harness’s core strength has always been… that it has somewhat of an opinionated way of doing things. That’s actually coming into vogue with the platform engineering movement.” Upcoming Harness features include agentic workflows for release orchestration and AI/MLOps beta modules.

Chris Williams, Head of DevSecOps at Takeda, at a Harness event, said, “[Given] the importance of responsible AI, I look at the SDLC and CI/CD specifically as a great enforcement point.” This underscores agents’ role in enforcing policies amid AI code proliferation.

Challenges in the Agent Era

Jason Andersen of Moor Insights & Strategy cautioned, “With agent-based products, it’s not just about features, but also how adoption and feedback drive improvements in AI accuracy and efficiency. I’m interested in how much effort will be required to modify and ground the prebuilt GitLab agents for [a specific] environment.” Environment-specific tweaks and transitional human oversight remain hurdles, as Blakely noted: “They’re not ready for that yet, but I’m a big believer that this whole process could be automated.”

Mitch Ashley, VP at Futurum Group, affirmed, “Success using AI agents will be heavily influenced by the context with which they operate. GitLab Duo Agent Platform leverages GitLab’s system of record to provide AI agents with deep project context for orchestrated, asynchronous collaboration.” Harness raised $240M led by Goldman Sachs in December 2025 to fuel AI expansion, valuing it at $5.5B, signaling investor confidence in post-code automation.

GitLab’s platform supports self-hosted models for sovereignty, with group-based controls for phased rollouts. As enterprises deploy, human-AI interplay will define viability, with both vendors converging on hybrid models per analysts.

Forward Momentum in DevSecOps

GitLab Duo Workflow private beta waitlist opened February 2025, promising visibility and control for agentic AI. Harness eyes library expansion using merged Traceable data for security patterns. Production proof will validate claims, as Volk concluded: “Instead of owning the entire platform like GitLab does, Harness gets its context from agent interaction. Which one is better? The proof lies in production deployments.”