absent-sapphire•2y ago
Hello please help
So I have created some intents on the flow and if there is a no match to answer from the knowledge base but for some reason it's not working and I dont know why. If I ask something which is a no match it answers it but if I ask something from the intents its says no match and the flow stops. You can see I have asked the question "I want to buy a property" - "Искам да купя имот" in Bulgarian and it cant find a match and flow stops but if it doesn't find a match shouldn't it go to knowledge base. If I ask a question in the beginning of the flow from the intents it recognize them it send it to the intent on the flow but if I ask a no match question and after that a question from the intents it just stops... Can you help please?
2 Replies
It seems like you're experiencing an issue where your Assistant is not correctly identifying and responding to intents after a 'no match' scenario. This could be due to the way your Assistant is structured in Voiceflow.
Voiceflow prioritizes responses in three levels: Intents, Knowledge Base, and Global No Match. Intents are designed by you to trigger based on specific questions. If the Assistant can't find a match in the Intents, it will look into the Knowledge Base. If it still can't find an answer, it will resort to the Global No Match message.
In your case, it seems like the Assistant is not able to find the correct intent after a 'no match' scenario. This could be due to the way the intents are structured in your flow. It's important to ensure that the intents are clearly defined and that the Assistant can correctly identify them.
Also, the Global No Match setting could be affecting the flow. If you have set the Global No Match to 'static', the Assistant will respond with a predefined message when it can't find a match in the Intents or Knowledge Base. If you want the Assistant to generate a response using AI, you should set the Global No Match to 'generative'.
Remember, the Knowledge Base is an experimental feature leveraging Large Language Models (LLMs), so there is potential to generate misleading or false information. It's currently only available for use with your No Match fallbacks, and cannot be used on a specific response within your Assistant.
I hope this helps clarify how the Assistant prioritizes responses. If you need further assistance, I recommend reaching out to the community members who have experience with similar issues.
absent-sapphireOP•2y ago
I have tried all but still doesn't work