Hello everyone! What do you recommend for the following case? I want a tool that, based on something the client says while talking on the phone with the Vapi assistant, if my assistant can’t resolve it, sends a request to an endpoint in my service that internally runs an LLM to process the request. However, this can take some time, and I don’t want to keep the connection open. Is it possible for my service to return a 200 ACK, and then, once the LLM finishes processing, send the tool call result back to Vapi (to my assistant)? In this case, does the assistant keep talking, or can it wait until that response arrives?