Snowflake Launches OSI: A Universal Standard for Business Intelligence Data
Snowflake has launched a vendor-neutral initiative called Open Semantic Interchange (OSI). The project aims to create a universal standard for adding context to structured and unstructured business intelligence (BI) data. This move comes amidst the growing need for a common data language, especially with the increasing use of AI agents and the presence of multiple BI tools within organizations.
OSI aims to be a universal semantic data framework, enabling different organizations to share data across platforms using a common set of definitions. The OSI working group plans to base the format on YAML, a widely used configuration language. This choice ensures wide adoption and ease of use.
Snowflake, in collaboration with Salesforce and dbt Labs, is leading the OSI effort. Other notable partners include BlackRock, Mistral AI, and RelationalAI. The group is inviting more organizations to join, fostering a collective approach to standardizing data definitions.
The need for OSI is amplified by the current data landscape. Snowflake customers, for instance, use a median of five different BI tools. Each tool potentially defines concepts and formulas differently, leading to inconsistencies. The growing popularity of AI agents further exacerbates this issue, as they rely on accurate, standardized data to function effectively.
OSI seeks to address these challenges by providing a standard definition for metrics, allowing AI systems and BI tools to use the same calculations. BlackRock has already shown support by planning to integrate OSI with its Aladdin platform. The project is expected to transition to being managed by a vendor-neutral, independent standalone entity in the future. By creating an open source repository for the specification and converters, OSI aims to facilitate data sharing and interoperability across platforms.
