dbt Labs Offers New Agentic AI Features, Powered by Fusion

dbt Labs, a leader in standards for AI-ready structured data, is offering dbt Agents, a suite of intelligent AI assistants built into dbt and made accessible via the remote dbt MCP server—supercharging development, improving governance, and delivering trustworthy AI outcomes.

According to the company, these platform updates accelerate development, support cross-platform portability, and lay an important foundation for new analytics use cases. 

“Open standards and AI are fueling the next era of analytics, and the dbt Fusion engine is the bridge that data teams need to move toward that future,” said Tristan Handy, founder and CEO at dbt Labs. “Fusion delivers robust context, tools and error-correction mechanisms for both humans and agents. It is the enabler of next generation, AI-powered data infrastructure.”

These governed AI agents are powered by dbt’s context, to make analytics faster and smarter while preserving quality, trust, and governance. dbt Agents is built directly into the platform and includes:

• Developer agent: Explains logic, flags duplicates, validates, and authors/refactors from prompts in VS Code or dbt Studio for faster, safer shipping. 
• Discovery agent: Finds the right datasets and definitions, highlighting trusted sources for faster exploration.
• Observability agent: Monitors jobs, identifies likely root causes, and proposes fixes to reduce manual remediation work.
• Analyst agent: Built into dbt Insights, this agent answers questions about models, jobs, and metrics, dramatically accelerating the insight-generation process.

These agents bring AI into the heart of the Analytics Development Lifecycle, helping teams accelerate outcomes, improve quality, and maintain governance, ensuring AI delivers tangible business impact.

This structured context is now universally accessible to AI systems through the remote dbt MCP server, now Generally Available. It runs in the cloud, connecting AI tools to projects in dbt without local setup. Now, dbt’s context, tooling, and error-correction is accessible to model providers and IDEs like OpenAI, Anthropic, and Cursor for safer, more reliable AI systems. 

Fusion, now in Preview for eligible projects on BigQuery, Databricks, Snowflake and Redshift, is building on its supercharged developer experience capabilities and enabling customers to dramatically optimize compute spend, eliminate wasted cycles, and focus teams on innovation and faster insights delivery.

Furthermore, Fusion is powering more than just smarter orchestration; it is supporting evolving analytics use cases. dbt-powered pipelines can now create and manage Apache Iceberg tables in Snowflake and Databricks, laying the groundwork for easier adoption of open table formats and cross-platform portability, according to dbt Labs.

In addition, the dbt VS Code Extension, now in Preview, lets developers run Fusion locally for tighter inner loops and parity with production environments, while dbt Insights, now Generally Available, uses Fusion’s language server bringing definitions, lineage, cost, performance, and reliability into one place for faster, smarter decisions.

Lastly, dbt Labs also announced that MetricFlow is now fully open source, with an Apache 2.0 license. This reaffirms dbt Labs’ commitment to the Open Semantic Interchange (OSI) initiative. By standardizing metrics and semantics across tools, organizations can ensure consistent, trusted outcomes across analytics and AI workflows, the company said.

For more information about this news, visit www.getdbt.com.