Per-agent overrides for inbox, AI provider, prompt, and runtime UX.
How quickly the AI's response unrolls in the chat. Slower feels more thoughtful; faster reduces wait perception.
The "we're working on it" beat between the AI seeing the input and starting to respond. Editable copy + animation style.
The little phrases sprinkled through the runtime. Each can be Whitelabel-rewritten via the inline AI editor.