Hi!
I'm building a conversational agent with Dialogflow CX for my company but an issue about access to information has been raised in the last meetings. Basically, they want the chatbot to answer to questions about certain topics only to selected people while not replying to not authrized users.
Additional info: Both of their data store are connected to a google cloud storage bucket. The chatbot is integrated to Google Chat.
I've tried using the info provided by the documentation (https://cloud.google.com/generative-ai-app-builder/docs/data-source-access-control#acl-storage-unstr...) but it doesn't seem to work.
What I've tried so far:
- selecting the identity provider on the Vertex AI Platform
- uploaded the documents and the metadata file with also the acl info in a Cloud Bucket
- created a new data store with ACL
- connected the data store to the metadata file
Firstly, I wasn't able to see the documents imported in the data store. It says that's because I created a data store with ACL. The data store shows that there are documents in it but I cannot see them.
Then, I tried to test the chatbot but it cannot answer any question (no match). I checked the logs and it says that it cannot find any document related to the question, an empty search.
My question is: is it possible to implement ACL in a chatbot made with dialogflow cx integrated with google chat? If so, what did I do wrong and what do I have to do in order to implement ACL?
Thank you!
LinkedIn
Twitter
