​ To learn more, see the sources I considered: [[1] custom function issue where static text is said after function received the response](https://discord.com/channels/1186424376300679209/1460607322358546567) [[2] Conversation Node](https://docs.retellai.com/build/conversation-flow/conversation-node) ​

### Answer This behavior has been reported by other users as well. [(1) ]() The expected flow is that static text plays during function execution, but sometimes the static text plays after the function completes instead. Here are some workarounds that have helped others: 1. **Use a faster model**: Set the function node to use GPT-4.1 Fast to reduce latency. 2. **Disable "Speak During Execution" on the function node**: Instead, add a separate conversation node with "Skip Response" enabled before the function node to speak your filler text (e.g., "Let me check that for you") . This separates the speech from the function execution timing. 3. **Check your model selection**: Ensure both the conversation node and function node are using fast models like GPT-4.1 or GPT-4.1 Fast. The conversation node's "Skip Response" setting means the transition happens when the agent finishes speaking, then moves to the next node. [(2) ]() However, the timing interaction with subsequent function nodes can sometimes cause the speech to be delayed. If you continue experiencing this issue, sharing your call ID would help the team investigate further.

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer I understand the frustration—you've set up the flow correctly with a conversation node (static text + Skip Response) before the function node, but the static text still doesn't play until after the function executes. Based on the documentation, when "Skip Response" is enabled on a conversation node, the transition should happen when the agent finishes speaking, then move to the next node. [(1) ]() However, this timing issue with function nodes appears to be a known behavior that others have encountered. Here are additional steps to try: 1. Verify model speed on both nodes Change to a faster LLM and faster TTS to reduce response time. [(2) ]() Ensure both your conversation node and function node are set to fast models like GPT-4.1 Fast. 2. Consider using Node Functions instead of Function Nodes Rather than using a separate function node, you can add functions directly to a conversation node. [(3) ]() With node functions, the agent can continue talking while a function executes, and function results are available to the LLM for generating follow-up responses. This approach keeps the conversation flowing while the function runs in the background. 3. Share your call ID If the issue persists, sharing a specific call ID would help the team investigate the timing behavior in your specific flow. The distinction is important: Function Nodes execute deterministically on node entry, while Node Functions are LLM-driven and allow full dialogue with functions available simultaneously.

​ To learn more, see the sources I considered: [[1] Conversation Node](https://docs.retellai.com/build/conversation-flow/conversation-node) [[2] Performance-related configurations and noticeable latency during live calls](https://community.retellai.com/t/performance-related-configurations-and-noticeable-latency-during-live-calls/747) [[3] Node Functions](https://docs.retellai.com/build/conversation-flow/conversation-node-functions) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.