The speak during/after execution options don't show in subagents because that UI is only available on standalone function nodes it's a Retell limitation at the subagent level. Are you building this from scratch or already have something running?