Delivering reliable AI with the dbt Semantic Layer and dbt MCP Server

Описание: As teams race to adopt LLMs and agents, many discover the same challenge: AI systems can’t perform well without structured, governed, and consistent context. Critical knowledge—from metrics and definitions to policies, data relationships, and operational rules—often lives in dashboards, documentation, or scattered SQL. Without a shared, authoritative layer of meaning, AI outputs become unreliable and leads to hallucinations, inconsistent answers, and costly rework.

In this virtual event, experts from dbt Labs and phData will walk through how the dbt Semantic Layer and dbt MCP Server combine to deliver the reliable context AI systems require. You’ll see how the dbt Semantic Layer centralizes and standardizes meaning—and how the dbt MCP Server extends that context directly to AI tools, enabling them to safely discover, query, and reason over trusted, governed data.

What you'll gain from watching:
A clear understanding of the AI context problem.
Learn how inconsistently defined data, duplicated logic, and ungoverned SQL make LLMs brittle—causing incorrect answers, unpredictable agent behavior, and unreliable reasoning. Discover how dbt helps teams align on meaning and provide AI systems with trustworthy context by giving AI access to the same governed knowledge your analytics tools rely on.

Demo: AI-ready data in action.
See a walkthrough of what’s possible with dbt today, including:
-How AI tools use the dbt MCP Server to discover available models, metrics, and metadata
-How the agent can ask for context to understand attributes and navigate relationships
-How the dbt MCP Server enables AI models to generate governed SQL
-How queries can be executed safely using dbt MCP Servers’ built-in guardrails
-How consistent semantic context leads to more accurate, predictable, and cost-efficient AI behavior
-Plus a quick demo of an agentic chatbot interacting with OpenAI using dbt MCP

Guidance you can use immediately.
Leave with practical next steps for getting started with dbt modeling, the dbt Semantic Layer, and dbt MCP Server in your own environment.

Meet the speakers:
Dustin Dorsey, Data Engineering Director at phData
Dakota Kelley, Principal Solutions Architect at phData
Stephen Robb, Partner Solutions Architect at dbt Labs