Harness is releasing four new capabilities for its Continuous Delivery platform. These are Release Orchestration, AI-Powered Verification and Rollback, Database DevOps with Snowflake support, and Warehouse-Native Feature Management and Experimentation. The additions target a growing mismatch between how fast teams write code and how reliably they can release it.
AI coding assistants have fundamentally changed the pace of software development. Engineering teams generate more changes across more services than ever before, but their release processes have not kept up. Harness calls this the velocity paradox, and its own 2026 State of DevOps Modernization Report puts numbers to it: teams using AI coding tools most heavily release daily or more in 35 percent of cases, yet those same teams see the highest remediation rate for deployments (22 percent) and the longest mean time to recovery at 7.6 hours.
That gap between writing code and safely shipping it is precisely where Harness is investing. Earlier this month, Harness already expanded its platform with AI Security and Secure AI Coding modules to address vulnerabilities introduced by AI-generated code. The four capabilities announced today extend that platform further, covering the full arc of a release rather than any single stage. In February, Harness also launched Artifact Registry with integrated AI security checks, reducing the time between vulnerability diagnosis and remediation by 83 percent.
Four capabilities, one answer
Release Orchestration replaces the Slack threads and spreadsheets that still coordinate most multi-team releases. Services and supporting teams move through shared orchestration logic with consistent controls and sequencing. The company opted for a native approach to Continuous Delivery rather than one that was bolted on separately.
AI-Powered Verification and Rollback connects to existing observability stacks, automatically identifies which signals matter per release, and decides in real time whether a rollout should proceed, pause, or reverse.
Database DevOps now includes Snowflake support, bringing schema changes into the same pipeline as application code with shared auditability. Should a rollback be needed, application and database schema can be rolled back together. The fourth capability, Warehouse-Native Feature Management and Experimentation, lets teams progressively expose features and measure business impact directly in their data warehouse, skipping ETL pipelines and shadow infrastructure.