Speech biases are influencing the speech to text recognition. So you can for example add here names and other slots etc.
For example you have a Hotel named King Thegeorge you can add it. So google will try to recognize it and deliver the String to you.
but one question, how you use the rest of the inspected input? Do you found a way to give him the possible intents and got them in the request after your response?
you can use them like that (slots is an array, message a string):
function answerWithMessage(message,slots){
let token = {
"state":null,
"data":{}
};
let jsonResponse = {
conversationToken: JSON.stringify(token),
expectUserResponse: true,
expectedInputs: [
{
inputPrompt: {
initialPrompts: [
{
textToSpeech: message
}
],
noInputPrompts: []
},
possibleIntents: [
{
intent: "actions.intent.TEXT"
}
],
speechBiasingHints: slots
}
]
};
return JSON.stringify(jsonResponse,null, 4);
}
if you create the json response by yourself