----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Glossary]() * [Orchestration Models]() * [Introduction]() * [List Assistants]() -----------------------------

### Answer Vapi gives you the ability to control the backchanneling behavior during conversations. While the specific terms like "yay" or "yeah" are not mentioned, the system allows for customization of backchannel cues. Vapi uses a proprietary fusion audio text model to determine the best moment to backchannel and to decide which backchannel cue is most appropriate to use [(1)]() . If you want to avoid enthusiastic affirmations, you might consider customizing the cues to more neutral responses like "mm-hmm" or "I see." Currently, Vapi does not provide a built-in switch to toggle between different emotional tones of backchanneling, but the flexibility of the system should allow you to define the type of responses that fit your needs. Additionally, the implementation of sentiment analysis does not automatically adjust backchannel responses in Vapi; these need to be explicitly configured. For reducing perceived latency, maintaining some form of backchanneling, even neutral, is beneficial as you noted. If you need more detailed guidance on how to implement these customizations, the Vapi documentation on orchestration models would be a good resource. Should you have further inquiries or need specific implementation assistance, diving deeper into the backchanneling section under the How Vapi Works documentation would be beneficial.