dbt Agents and the Remote MCP Server Are Shaping Trusted AI Analytics

Analytics is entering a new era, driven by AI-powered agents and robust governance frameworks. dbt Labs recently unveiled the remote dbt MCP Server alongside dbt Agents, a suite of secure, task-specific AI agents that leverage dbt’s structured context layer. These innovations are set to accelerate analytics workflows while maintaining strict data quality and governance.

Structured Context: The Backbone of Reliable AI

For AI agents to deliver accurate insights, they must operate within a clearly defined and accessible structured context. Without this foundation, AI can produce inconsistent or unreliable results. dbt’s long-standing approach including defining data lineage, tests, and semantics helps to ensure that AI has trustworthy rules to follow.

The newly released remote dbt MCP Server makes this structured context available as a secure, standardized interface. Now, both in-house and third-party AI tools can interact with governed analytics projects without cumbersome local setups or custom integrations, streamlining access and collaboration across environments.

Remote dbt MCP Server: Standardizing Agentic Analytics

With general availability in both remote and local deployments, the dbt MCP Server turns structured analytics context into a universal endpoint. This enables trusted AI partners like OpenAI and Anthropic to safely tap into governed models, metrics, and lineage, making analytics workflows more explainable and auditable.

• OAuth authentication boosts security and audit controls for local use, with remote OAuth support coming soon.
  
  
• dbt Fusion engine provides advanced compiler, diagnostics, and metadata tools, allowing AI agents to validate transformations and catch issues early.

These features give teams the confidence to adopt agentic workflows, knowing access is tightly governed and operations are dependable.

dbt Agents: AI Built for Analytics Teams

Building on the MCP Server, dbt Agents are prebuilt, governed AI agents that enhance analytics development, management, and consumption. They operate within dbt’s trusted context, ensuring every action aligns with best practices and governance policies.

• Analyst Agent (beta): Answers natural language questions using governed results by generating and executing SQL from dbt models. It references metric definitions and lineage for accuracy.
  
  
• Discovery Agent (beta): Helps users find datasets and metrics using natural language, surfacing reliable definitions and sources for safer self-service analytics.
  
  
• Observability Agent (coming soon): Monitors jobs, diagnoses root causes, and accelerates issue resolution to minimize operational disruptions.
  
  
• Developer Agent (coming soon): Provides model logic explanations, predicts impacts, and validates code directly in development environments, empowering engineers to deliver quality changes rapidly.

These agents handle routine analytics tasks, freeing up engineers for strategic work and giving business users easier access to trustworthy data.

dbt: The Standard for AI-Driven Analytics

dbt’s structured context, exposed through the MCP Server, makes it the foundation for reliable agentic analytics. With over 900 teams already experimenting with conversational and AI-driven workflows, dbt is demonstrating that explainable and governed AI is crucial for scaling analytics across organizations.

Innovations like open-source MetricFlow and dbt Fusion reinforce this foundation by providing consistent metrics and reliable code validation. As AI agents become more advanced, dbt’s commitment to open standards and comprehensive metadata will only boost their effectiveness and trustworthiness.

Towards Autonomous, Scalable Analytics

The introduction of dbt Agents marks a shift from manual analytics requests to proactive, automated workflows. As these AI agents mature, they will enable smaller teams to manage larger datasets and more complex analytics pipelines, all without sacrificing governance or data quality.

dbt Labs is dedicated to expanding agent capabilities, deepening AI integrations, and enriching structured context with even more metadata and intelligence. The vision: empower data engineers to innovate and collaborate at scale, while AI agents handle repetitive, operational tasks.

If you’re ready to modernize your analytics, enable the  the remote dbt MCP Server, try the Fusion MCP tool, and request access to dbt Agents waitlist to help shape the future of trusted, AI-powered analytics in your organization.

Source: dbt Labs Blog