Information Technology
dbt Labs Affirms Commitment to Open Semantic Interchange by Open Sourcing MetricFlow
Inconsistent metrics and fragmented definitions have become a persistent problem as enterprises rush to deploy AI, leading to diminished trust and slow adoption. Because of this, there is a massive need for a single standard for metrics that every tool and agent can rely on along with an engine that renders a metric into its provably-correct calculation. MetricFlow is the core engine that compiles metric definitions into the code that computes them, and, unlike text-to-SQL methods, that computation is explainable and reliable every time.
MetricFlow, which has powered the dbt Semantic Layer following the company's acquisition of Transform in 2023 , uses information from semantic model and metric YAML configurations to construct and run SQL in a user's data platform, providing governed metrics. By open sourcing MetricFlow under the Apache 2.0 license, dbt Labs is providing the community with a transparent, extensible engine that will enable AI agents to leverage trusted metric definitions for governed conversational analytics, ensuring teams get consistent results across tools and clouds for scale.
"dbt is rooted in our open source DNA. This transition to open sourcing MetricFlow will unlock new opportunities for data practitioners to deliver huge value for their organizations," said Ryan Segar , Chief Product Officer of dbt Labs. "Metrics drift across tools and trust erodes if there isn't a single source of truth, and with 90% of analysts agreeing that they desperately need more efficient tools to meet business demands, open sourcing this engine greatly diminishes the cadence of re-work while improving organizational trust in data."
dbt Labs' recent commitment to the OSI initiative, led by Snowflake , the AI Data Cloud Company, and other industry leaders including Salesforce and Sigma, focuses on solving the costly impact of nonstandardized data definitions.
"The OSI initiative aims to standardize how semantic metadata is defined and shared, and dbt Labs' change of the MetricFlow license to Apache 2.0 is an important step in this mission," said Josh Klahr , Director of Analytics Product Management at Snowflake. "Fragmented data definitions are one of the largest barriers to AI adoption, and we believe that MetricFlow can serve as a key component in helping us achieve our vision of providing a shared set of analytic metadata that will propel trusted data forward."
The collaboration acknowledges the pitfalls that come with proprietary semantic standards and offers a solution to directly target the bottleneck holding organizations back from achieving lofty AI-driven goals. MetricFlow will be a central component in realizing the vision of the OSI initiative.
"Defining metrics in MetricFlow is crucial for a unified source of truth," said Rob Vicker , Data Analytics Architecture Director at EMS Insurance. "However, BI and AI tools often interpret metrics in their own ways. By leveraging open-source MetricFlow, OSI can ensure every tool is consistent in the consumption of metrics. This saves analysts time, streamlines audits, and gives users the flexibility to access data as they prefer, without constant code updates."
For more information, visit https://www.getdbt.com/blog/open-source-metricflow-governed-metrics .
Since 2016, dbt Labs has been on a mission to help data practitioners create and disseminate organizational knowledge. dbt is the standard for AI-ready structured data. Powered by the dbt Fusion engine, it unlocks the performance, context, and trust that organizations need to scale analytics in the era of AI. Globally, more than 60,000 data teams use dbt, including those at Siemens, Roche and Condé Nast.
Learn more at getdbt.com , and follow dbt Labs on LinkedIn , X , Instagram , and YouTube .
Logo - https://mma.prnewswire.com/media/2779190/dbt_Labs_Logo.jpg
View original content:https://www.prnewswire.co.uk/news-releases/dbt-labs-affirms-commitment-to-open-semantic-interchange-by-open-sourcing-metricflow-302582866.html