Vertex AI search and Conversation

[VishalBulbule]

I am exploring Vertex AI search and conversation to build some usecase within my organization. I am looking for more technical details around vertex AI search and conversation . I have below queries
1. Does Vertex AI search and conversation use RAG for data retrieval?

2. If Yes how RAG works in Vertex AI search and conversation , is it within datastore ?

3. When we create datastore , does it create embedding and load in vector database at the backend?

Please someone help with relevant document/answers.

@xavidop help here please.

[xavidop]

Hi @VishalBulbule ,

Yes, vertex AI search and conversations is an no-codeRAG platform. When you create a data store it automatically generates the embeddings  using an LLM and store them in a vector database. Then you can use that data store in any Vertex AI search & conversation applications: search, recommendation or chat app.

Best,

Xavi

[VishalBulbule]

Thanks @xavidop  I am having same understanding. Thanks for confirming !

[xavidop]

cool! if you want, you can close this post and setting my comment as the solution!

[gdsgcp]

Hi Guys, is it advisable to use vertex AI datastore as backend database? we are in our  poc stage, and looking to upload a small csv file containing application/software directory.