Hi! I just noticed that, when I added new assistant, it seems that my previous assistants and the new added assistant have the same memory. They have different prompts and different knowledge base files. Upon testing, the new added assistant said something that it should not. It said something that is related to the other assistant. How to fix this?

@Gelo Hello! It sounds like there might be an issue with the configuration of your assistants where they are inadvertently sharing information. To resolve this, you should check the following: 1. Ensure that each assistant has its own unique memory space and is not referencing a shared memory. 2. Verify that the knowledge base files are correctly assigned to each assistant and that there are no overlaps. 3. Check if there are any settings or configurations that might have been mistakenly copied from one assistant to another. If you've already checked these and the issue persists, it might be a more technical problem that requires further investigation. In that case, please reach out to support for more detailed assistance. Would you like me to guide you on how to contact support or is there anything else I can assist you with right now?

How can I know if they are not referencing a shared memory? @Vapi

@Gelo To ensure that your assistants are not referencing a shared memory, you can perform the following checks: 1. **Configuration Files**: Review the configuration files for each assistant to confirm that they point to separate memory storage locations or databases. 2. **Initialization Parameters**: When initializing each assistant, make sure that any parameters related to memory or state management are unique to each instance. 3. **Knowledge Base Assignment**: Double-check that the knowledge base files are correctly assigned and that there is no accidental sharing of files between assistants. 4. **Isolation Test**: Conduct a test by providing inputs that only one assistant should know based on its knowledge base. If another assistant responds with that knowledge, it indicates a memory sharing issue. 5. **Documentation Review**: Consult the documentation of the platform or framework you are using for creating the assistants to understand how memory management is handled. 6. **Platform Support**: If you're using a third-party platform, check their support resources or contact their support team for guidance on memory isolation between assistants. If after these checks you're still unsure or the issue remains unresolved, it may be necessary to reach out to the platform's support team for further assistance. Would you like help with contacting support, or is there anything else specific you'd like to know?

can you guide me step by step on each check? @User

Hey @Gelo To help track down this issue, could you share: - The call ID - When exactly this happened (the timestamp) - What response you expected to get - What response you actually got instead This would really help me figure out what went wrong!

Hi @Shubham Bajaj The call ID: (I can't find the call ID because we already called a lot so I can't find it in our call logs) When exactly this happened (the timestamp): since I can't find the call ID, I can't provide the timestamp What response you expected to get: I'm expecting for the assistant to continue the conversation What response you actually got instead: The assistant responded "Hello, this is Mary from Mary's Dental. How can I assist you today?" which it should not because we already created a different assistant and that response was nowhere to be found in our prompt and knowledge base. I think this is because before creating the system prompt for our current assistant, we used the template provided by VAPI for appointment setters and it may be got the information from there, that's why the AI behaved like that.

@Gelo just ping me in case your require further help.