The problem and solution

Today there's no clean and governed way to get a Collibra Knowledge Graph into the context format a given platform needs.
Every integration ends up as custom scripting or manual exports that break easily. The result: point-in-time exports go stale (semantic or metric drift), and LLMs and copilots hallucinate because they don't have trusted business definitions to work from.

The Governed Context Compiler (Flexible Semantic Blueprint) private preview is a context engine that maps the shape of your Collibra Knowledge Graph to the structure of the schema you're targeting.

Through a visual UI or a key-based YAML config, you describe how to traverse your graph and what to pull at each step. The engine does the rest—fetching live data and post-processing it into the exact machine-readable format the downstream platform expects, whether that's Snowflake Semantic Views, OSI, Databricks Metric Views, or an AI agent.

The tester profile

We are looking for hands-on preview participants who occupy one (or more) of the following roles. If your day-to-day involves bridging the gap between data governance and technical execution, we want you!

• Governance Manager / Data Steward
  • Your Role in This Preview: You will use our visual Mapping UI to configure the relationship between your specific Collibra operating model and target output schemas.
  • Your Core Need: A no-code interface to define, test, and maintain semantic mappings without writing custom integration scripts.
• Data Engineer
  • Your Role in This Preview: You will consume the generated outputs (YAML files, API responses) to bootstrap or validate semantic layers in platforms like Snowflake, Databricks
  • Your Core Need: Accurate, machine-readable output in the exact format your platform expects, available on demand via API.
• AI Engineer / Agent Developer
  • Your Role in This Preview: You will programmatically query Collibra’s governed context (via REST API or MCP tools) to ground AI agents with trusted business data.
  • Your Core Need: A stable, low-latency programmatic endpoint that returns a governed semantic blueprint for any given asset.

How the preview works

Selected participants will be granted access to this preview feature in their Collibra non production environments. The Collibra Product team will provide test activities that will help with initial configuration and facilitate feature exploration via this portal, the Preview Platform. 

To maintain a spot in the program, participants are required to share feedback actively: submitting satisfaction scores, logging feedback tickets and completing program surveys.