#general
Sep 15September 16, 2026
hello there, I wanted to ask question related with knowledge base
• I have ingested the knowledge through the connector.
• the docs already indexed.
• The agent plug only the knowledge base because I assume docs already indexed
but when I inquiry the agent about the recent incident (which should be exist in knowledge base), it cant find anything (?)
However, if I plug the connector, the agent was able to fetch through the connector.
am I understand the concept of knowledge base correctly or I am missing something here that makes the agent unable to see the indexed docs?
I haven't checked to see if it's been changed, but is anyone aware if the SharePoint knowledge connector still grabs all pages, published and draft?
New blog post alert :speech_balloon:
We built a tool internally, that fixes Archestra bugs through autonomous sessions launched from Slack. While it works well for small fixes, it touches internal data and has ability to create PRs - basically has a full complex of Lethal Trifecta. In this post we share how we made sure this agent is secure and easy to use. Spoiler - everything starts with dropping a :crab: in Slack.
Read-only live mirror of Archestra.AI Slack
👋Join the discussion withAI enthusiasts!hello there, I wanted to ask question related with knowledge base
• I have ingested the knowledge through the connector.
• the docs already indexed.
• The agent plug only the knowledge base because I assume docs already indexed
but when I inquiry the agent about the recent incident (which should be exist in knowledge base), it cant find anything (?)
However, if I plug the connector, the agent was able to fetch through the connector.
am I understand the concept of knowledge base correctly or I am missing something here that makes the agent unable to see the indexed docs?
3 replies
hello! 👋
which version are you running? v1.3.61 and v1.4.0-beta.10 (both released yesterday) made some improvements to how agents see/are made aware of available knowledge sources.
Can you try upgrading and let me know if you are still hitting this issue? 🙂
we are on 1.3.59 now
alright, we will try to test the latest release
👍1
seems like it is intended that connector must be exist together with knowledge bases from the codebase (?) so the agent can retrieve the knowledge
I understand the reason from design perspective that since connector can pull data from same source but different target (e.g github repo)
but i wonder why after ingestion why cant we just query from knowledge bases straight away since it is already indexed