by default the assistant starts speaking as soon as the call connects. To delay that so your server URL has time to return the customer info, you can: 1. Go to your assistant settings in the dashboard. 2. Under Behavior → First Message, disable “Start automatically” (or clear the “First message” field). 3. Then, from your server URL, include a short initial message in the response — once that returns, you can trigger the assistant to start speaking using the conversationUpdate event or API call. Docs: https://docs.vapi.ai/server-url/events This way, Vapi waits for your server to send data before the assistant speaks, ensuring {{customer.name}} etc. are ready to use.