Databricks launches AI agent for legacy SQL migration

Databricks is expanding its Lakebridge toolkit by introducing an agentic code conversion feature designed to help organizations migrate from legacy data warehouses. This new capability uses Genie Code to rewrite complex SQL scripts, allowing customers to transition their workloads to the Databricks lakehouse environment with higher efficiency and less manual intervention.

Advanced Automation for Complex Code Translation

The core of this update is the agentic code converter, a system that utilizes AI subagents to manage the heavy lifting of migration projects. These agents perform a variety of tasks including deep analysis of source code and the parallel conversion of multiple files. They also validate translated SQL and can autonomously retry sections that fail during the initial pass. This iterative approach is a significant step forward from traditional methods that often require human developers to step in when software hits a wall.

By allowing developers to set specific migration rules for unique enterprise SQL structures, the tool provides a level of customization that previous automated systems lacked. The Lakebridge suite already offers several transpilation engines, such as the pattern-based BladeBridge technology and the compiler-based Morpheus engine. However, the addition of agentic AI introduces a reasoning layer that these older technologies do not possess. This reasoning is vital for moving beyond simple syntax mapping and into the realm of complex logic.

Traditional transpilers like Morpheus are excellent at handling standard syntax mapping. They easily manage date functions and basic join commands. Problems arise when these tools encounter control-flow reasoning, cursors, or dynamic SQL that is generated at runtime. These complex elements often differ significantly across platforms like Oracle or Teradata. Industry experts note that these difficult sections usually represent about 15 percent of a codebase but consume the vast majority of manual labor during a migration.

The agentic converter is specifically built to tackle these high-effort areas. It handles stored procedures and proprietary functions by understanding the context of the code rather than just the literal text. This contextual awareness improves the quality of the first conversion attempt. For IT leadership, this improvement means shorter project timelines and a notable reduction in the specialized labor costs associated with fixing broken scripts.

Implementation Challenges and Current Limitations

While the potential for time savings is high, the technology is currently in its beta phase, which brings certain operational constraints. Organizations must manage their expectations regarding how quickly they can move entire systems. For instance, the current version of the tool limits conversion jobs to batches of 300 files at a time. This means that a large enterprise with thousands of stored procedures will still need to plan for multiple runs and include manual checkpoints in their workflow.

One way Databricks addresses these hurdles is through a new feature for custom migration skills. This allows engineers to turn one-off manual fixes into reusable rules that the AI can apply to future files. While this helps codify knowledge, the consistency of the agentic system is still being evaluated as the product moves toward a full release. Developers will need to monitor the system closely to ensure that the AI remains accurate across different types of legacy scripts.

It is also important to recognize that SQL conversion is only a single piece of the modernization puzzle. Even if the AI rewrites every script perfectly, teams still have to reconcile migrated data against their live production systems. They must also sequence complex dependencies and secure business approval before shutting down old hardware. Because the current Databricks strategy focuses primarily on the code rewrite, other phases like data reconciliation remain part of future development cycles.

The overall migration timeline may not shrink as much as some hope because of these other necessary steps. The tool effectively compresses the time spent on the SQL rewrite portion of the project, but the surrounding logistics still require careful human management. Organizations should view this as a powerful assistant that removes a specific bottleneck rather than a completely hands-off solution for digital transformation.

Competitive Landscape in AI Driven Migration

Databricks is participating in a broader industry trend where data platform providers use AI to lower the barrier for new customers. The competition for enterprise data is fierce, and making it easier for a company to leave a rival platform is a key strategy. By reducing the friction of moving code, these vendors hope to capture market share from established legacy providers who have long benefited from the difficulty of switching platforms.

Snowflake offers a competing tool known as SnowConvert. Like the new Databricks offering, it uses automated agents to help businesses move data warehouses, business intelligence reports, and underlying code to its own cloud platform. This reflects a growing standard in the industry where the “moat” created by complex legacy code is being bridged by generative AI and agentic workflows.

Google Cloud also provides similar capabilities through its Database Migration Service. By using Gemini-powered logic, Google enables automated workflows for schema and code conversion. This creates an environment where CIOs have multiple options for moving their workloads to the cloud. The choice often comes down to which tool handles the specific quirks of their legacy environment with the least amount of errors.

As these tools mature, the focus will likely shift from simple code translation to full lifecycle automation. This would include the data reconciliation and dependency mapping that currently remain manual tasks. For now, the move toward agentic converters represents a major shift in how technical debt is managed. Instead of spending months manually translating old scripts, developers can act as supervisors for AI agents that do the bulk of the work in a fraction of the time.