Likely Causes of the Issue 1️⃣ Vapi is not properly mapping the webhook response data → The assistant might still be using internal placeholders instead of your actual response. 2️⃣ Inconsistent parameter names between your webhook response and the assistant’s tool call → Example: caller_i_d vs. callerID. 3️⃣ Make.com isn’t returning the correct JSON structure → The assistant might not be correctly reading or parsing the webhook response. 4️⃣ The response from Make.com is delayed → If Make.com takes too long, the assistant might fall back to default placeholders. 🔹 Step-by-Step Debugging Guide ✅ Step 1: Check the Webhook Logs in Make.com Go to your Make.com scenario → Open the webhook module that receives the assistant's request. View the Incoming Data → Does it correctly contain destination and recipientName? Expected Example: json Copier { "destination": "+1405818XXXX", "recipientName": "Eric Huggins" } If the assistant is calling with placeholders like "Ricky Bobby’s phone number," it means Make.com isn't sending the correct response. ✅ Step 2: Validate the Webhook Response Format Check your HTTP Response Module in Make.com Open the HTTP "Make a Request" module that responds to Vapi. Ensure the JSON is structured exactly as Vapi expects. The Correct Response Format Should Be: json Copier { "destination": { "type": "number", "number": "+1405818XXXX", "recipientName": "Eric Huggins" } } Make sure: destination is an object (not a string). The key names exactly match what Vapi expects. recipientName is included. ✅ Step 3: Ensure the Assistant Uses the Webhook Response Go to Vapi’s Assistant Settings Check if there’s any hardcoded logic that could be overriding the webhook response. Look for assistant behaviors where it tries to infer "Ricky Bobby’s phone number" instead of using the webhook data. Modify the Assistant’s System Prompt to Force It to Use the Webhook Response Try updating the system prompt with this instruction: