I think vapi has something in the UI called success evaluation which allows you to write prompts to grade the success criteria of your call But what it doesn't have I think is 1. Grade the AI response through cosine similarity 2. A system as robust as bland pathways. Bland pathway allows you breakdown your prompts & guide the AI on which prompt it should use during the conversation based on what condition rules are passed or not. This makes it easier to find and understand what went wrong in the conversation & what to fix prompting wise especially when combining it with the testing features Bland has. Now pathways still comes with pro and cons & vapi does have something similar called blocks though it's beta right now & nowhere near as robust as Bland pathways