In a previous blog, we used Tavro’s Open-Source Agent BizOps Platform to generate an AI Use Case and Agent to originate sub-prime customers with less than 650 credit scores. How do we integrate this back to the data governance program?
Please visit the Tavro GitHub repo for a detailed how-to on the Collibra integration.
In this blog, we demonstrate a use case to integrate Tavro with Collibra’s MCP Server:
- Prototyped AI Agent using Tavro
To recap from the earlier blog, we prototyped the AI agent in Tavro.
- Configured Collibra MCP Server from Tavro
From the Tavro Admin portal, we configured the connection to the Collibra MCP Server and available MCP tools including create_asset, which we will user later in this blog.
We added Anthropic (Claude) as the provider type and entered the API key.
- Identified Critical Data Elements for the AI Use Case
We prompted Tavro to provide five Critical Data Elements (CDEs) for the AI use case.
Tavro generated the CDEs.
- Created CDE’s in Collibra using Collibra MCP Server
We then asked Tavro to create the critical data elements in Collibra in the Banking domain. If you do not specify the domain, the LLM will make its own intelligent determination.
The CDE’s are automatically generated in the Banking Domain in Collibra.
Want to discuss further? Join our Slack channel.
