Sometimes, VAPI does not ask the user all the questions provided in the prompt. Instead, it assumes values for the function parameters and passes them to the backend API. Occasionally, VAPI uses example values, and other times it makes assumptions. I have already included the following warning in the assistant instructions. Please let me know how I can fix this. [Warning] Never assume value for function or tool parameters. Never pass the example value to the function or tool parameters. Call ID: 4866818b-ff83-4e6b-8e62-29735b8e8cff

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

Call ID: 4866818b-ff83-4e6b-8e62-29735b8e8cff Timestamp: May 21, 2025, 15:14 (Got it from VAPI Dashboard) The AI should have asked the caller which hospital they were calling from, name of the caller and the message that needed to be passed to the backend API. I got the following information to the backend API - Message - I want to talk to the on-call doctor. (This is an example I provided to the assistant) - Callername - Hospital Caller (VAPI assumed this value) - Hospitalname - Atrium and Miami Valley South Hospital (VAPI assumed this value)

Hey Diana, I reviewed your Shared CalI ID logs. You're using GPT 4.0 mini with a model temperature of 0.7. I suggest you use GPT 4.1 or GPT 4.0 with a model temperature between 0 and 0.3 to better understand the data and craft responses as desired. Try 2-3 test calls and let me know how it goes.