Harness Inc., a software delivery startup that provides artificial intelligence tools for developers to update and monitor applications, said today it’s launching a new tool to automate database changes.

The new release, dubbed AI-Powered Database Migration Authoring, is part of the company’s Database DevOps platform, designed to deliver full-featured, enterprise-grade database structure and continuity automation, similar to what developers have come to expect from AI-assisted coding tools.

As a company steeped supporting developers throughout the application code lifecycle — from testing and integration in development to post-release support — Harness said it has observed that one of the biggest blockers slowing teams down is database change management.

Although developers can move fast when writing code, many organizations get bogged down when it’s time to upgrade their data backends. Legacy tools often make schema updates through SQL slow and cumbersome, with manual tracking and approval processes further delaying releases.

To address these slowdowns, Harness said it decided to apply the “vibe coding” concept to database migration, giving administrators AI-powered tools to handle testing, security and delivery for data changes more efficiently.

In a recent survey, Harness found that 63% of organizations ship code faster after adopting AI tools, yet 72% experienced at least one production incident due to AI-generated code. While AI has accelerated coding speed, it has also introduced new challenges for delivery. Harness believes that to truly speed up delivery, AI must extend across the entire development lifecycle, closing the gap for DevOps teams from code creation to the “last mile” of database automation and governance.

Harness believes to speed up delivery, AI must extend across the entire development lifecycle, closing the gap for DevOps teams building and maintaining code into the last-mile of database automation and governance.

Using AI-Powered Database Authoring, any developer can quickly describe data structure changes at the database level like an administrator. For example:

“Create a table named animals with columns for type, genus_species and common_name. Then add a table named birds that interfaces with animals and lost_and_found and adoption tables. Add rows for parakeet, cockatiel, cockatoo, canary, parrot, lovebird, finch and conure.”

From there Harness will generate a compliant, production-ready migration — complete with scripts for rollback, validation and Git integration. Before running, the AI automation examines current schema and policies, generates correct backward-compatible migration plan and validates the change for safety and compliance.

The result is a versioned, tested and fully auditable migration, much like the update process for application code.

The AI migration system has access to the current database structure and understands triggers, key-value pairs and relationships between tables. This enables it to prepare appropriate migration criteria based on natural language input. Database administrators retain human-in-the-loop oversight through policy enforcement and automated approvals.

“Harness gave us a truly out-of-the-box solution with features we couldn’t get from Liquibase Pro or a homegrown approach,” Daniel Gabriel, senior software engineer at cloud-based healthcare support company Athenahealth Inc. “We saved months of engineering effort and got more for less — with better governance, orchestration and visibility.”

Harness added that the new AI-driven automated database authoring capability is powered by the company’s Model Context Protocol server, which allows third-party AI models and agents — including Anthropic PBC’s Claude Desktop, Windsurf, Cursor or custom AI agents — to connect with Harness.

Image: alengo/Getty Images