https://retellai.com/?utm_source=discord&utm_medium=seo logo
Does knowledge base names matter
# ❓・ai-evy-support
b
@AI Evy Does knowledge base names matter? Can I sepcifically trigger an knowldge base retrival by name?
a
Thanks for tagging me @Bobster2002. I'll look through relevant sources and get back to you shortly.
### Answer Knowledge base names are primarily for organizational purposes—they help you identify and manage your knowledge bases (must be less than 40 characters) . [(1) ]() You cannot specifically trigger retrieval by name during a call. When a knowledge base is linked to an agent, the agent will automatically try to retrieve information from the knowledge base before responding—there's no need to change your prompt for it to trigger, as it will be done automatically for every response generation. [(2) ]() The retrieval process uses the transcript so far (prompt is not included) to find the most relevant chunks from the knowledge base and feed them to the LLM as context. If you need to link specific knowledge bases to an agent, you do so via
knowledge_base_ids
(an array of knowledge base IDs, not names) . [(3) ]() [(4) ]() So the name is just a label for your reference—actual retrieval is automatic and based on transcript relevance, not triggered by name.
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
​ To learn more, see the sources I considered: [[1] Get Knowledge Base](https://docs.retellai.com/api-references/get-knowledge-base) [[2] Knowledge Base](https://docs.retellai.com/build/knowledge-base) [[3] Update Conversation Flow](https://docs.retellai.com/api-references/update-conversation-flow) [[4] Get Retell LLM](https://docs.retellai.com/api-references/get-retell-llm) ​
b
@AI Evy why are knowledge base sometimes not triggered? Shouldn't it do it autommtically for every response generation?
a
Thanks for the follow-up, @Bobster2002. I'll get back to you soon.
### Answer Yes, when a knowledge base is linked to an agent, the agent will always try to retrieve information from the knowledge base before responding—it happens automatically for every response generation. [(1) ]() However, there are a few reasons why it might appear not to trigger: 1. Similarity Threshold Filtering The retrieval uses a similarity threshold (default: 0.6) . A higher setting gives you fewer, but more similar, matches. If the transcript content doesn't closely match the knowledge base content, chunks may not meet the threshold and won't be returned. 2. Timeout Issues There have been instances where the knowledge base timed out and then was not used. [(2) ]() This was identified as a root cause for erratic behavior in the past. 3. Agent Version Locking If your agent is published, it locks on the old KB state. You cannot edit a published agent, so you need to create a new version of it to pick up KB changes. [(3) ]() 4. KB Not Properly Attached Double-check that the knowledge base is actually attached to your agent—there have been cases where users thought it was attached but it wasn't. [(2) ]() You can configure retrieval settings via "Adjust KB Retrieval Chunks and Similarity" after connecting a knowledge base to your agent: - **Chunks to retrieve**: Max number of chunks (1-10, default: 3) - **Similarity Threshold**: How strict the matching is (default: 0.6) [(1) ]()
​ To learn more, see the sources I considered: [[1] Knowledge Base](https://docs.retellai.com/build/knowledge-base) [[2] Knowledge base is erratic](https://discord.com/channels/1186424376300679209/1306667196918792224) [[3] Knowledge Base not being picked up appropriately](https://discord.com/channels/1186424376300679209/1388800085638381599) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.