Dialogflow voice input too sensitive

[Sallamoss]

I would like to request help with the voice integration feature. I am currently building an agent for a business. The voice agent is extremely sensitive in regards to picking up sounds from everywhere. Even just shuffling can sometimes activate the barge-in. 

Is there any way to either decrease the sensitivity with which the agent picks up noises, or setup just a straight DTMF only setup through which voice interruptions might not be an option ?

One more question, is there a way to switch from Google voice recognition to Chatgpt whisper voice recognition?

[MarvinLlamas]

Hi @Sallamoss,

Welcome to Google Cloud Community!

It seems that your Dialogflow agent’s voice input is too sensitive, triggering unintended activations from background noise.

Here are potential ways that might help with your use case:

• Decrease Voice Input Sensitivity: You may want to adjust the end-of-speech sensitivity, or try enabling or disabling auto speech adaptation in your Dialogflow CX to reduce unintended sounds. Lowering the sensitivity helps minimize false triggers.
• Begin with Speech Adaptation: You may want to incorporate context phrases into your speech adaptation fields to refine the agent’s sensitivity to key phrases relevant to your user's interactions, thereby reducing the likelihood of the agent responding to unintended sounds.
• Explore DTMF as an alternative solution: You might want to enable DTMF (Dual-tone Multi-Frequency Signaling). It serves as a reliable alternative if voice input continues to pose challenges. This approach ensures that only intentional key presses trigger actions, providing enhanced control and completely eliminating background noise interference.

Furthermore, to answer your question regarding the switch from Google Voice Recognition to ChatGPT Whisper Voice Recognition, you may refer to the following documentation, which could help address your issue:

• Using Google Speech-to-Text API

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.

[Sallamoss]

Hello,
No unfortunately I did not.
Neither of the options was suitable and the voice activation happened way too early.
I was forced to switch to a Twilio integration. However, now I am facing more issues, which are related to Twilio. 

However, this issue got solved by that.