- 最新
- 投票最多
- 评论最多
I think that you may not need a chatbot at all. You could have the web site play a piece of recorded audio upon user landing (e.g.: using Amazon Polly if you want to use synthetic voices), then record the answer and use Amazon Transcribe if you want to turn it into text. Finally, you can play the farewell audio.
You could use Amazon Lex for the latter as well, but you'd need to have an intent with utterances that match whatever the user might say.
If you plan to extend the behavior of this very basic system to, for example, fulfill tasks for the user, I'd recommend using Amazon Lex (however, you will still need to trigger the initial audio from the app, as Lex will only respond to user interactions). In that case, you'd need to know what the user might answer and design intents for those tasks that are triggered by utterances that match the user's words.
相关内容
AWS 官方已更新 2 年前- AWS 官方已更新 2 年前
Are there any other interactions or anything that needs to happen in the backend upon the user's answer? or do you only need to record a spoken answer?
Only need to record a spoken answer. I’m so new to this. Any suggestions would be appreciated.