Advanced users of APIAI know that it was a little painful to work with input and output contexts
In simple words the conversation context structure needed a lot of thinking and setting of contexts.
With the new feature of "follow up intents" APIAI made it really easy and smarter to manage the conversations easily and in a well structured manner.
You can keep the user engaged in a structured talk without complex context settings.
It makes the talk really smart and the most close NLP based conversation.
Another important update is the opening up of all preintegrated domains for free. This is a huge feature. The agents for flight booking, restaurants, languages, alarms, smarthome. It opens a wide range of ideas to be explored.
Before I close this talk, let us solve very tricky and simple solution for a challenge.
Punctuation symbols are not shown in responses when posted via HTTP APIs for the agents.
e.g. if the speech response contains a comma or a full stop, it is not displayed in the final speech response.
Solution for this issue is : Once you get the speech as input in the POST in a parameter P; during the final response/output show the value as P.original. It will show the original speech.
If you do not use this trick, the speech might be without the punctuation symbols.
In simple words the conversation context structure needed a lot of thinking and setting of contexts.
With the new feature of "follow up intents" APIAI made it really easy and smarter to manage the conversations easily and in a well structured manner.
You can keep the user engaged in a structured talk without complex context settings.
It makes the talk really smart and the most close NLP based conversation.
Another important update is the opening up of all preintegrated domains for free. This is a huge feature. The agents for flight booking, restaurants, languages, alarms, smarthome. It opens a wide range of ideas to be explored.
Before I close this talk, let us solve very tricky and simple solution for a challenge.
Punctuation symbols are not shown in responses when posted via HTTP APIs for the agents.
e.g. if the speech response contains a comma or a full stop, it is not displayed in the final speech response.
Solution for this issue is : Once you get the speech as input in the POST in a parameter P; during the final response/output show the value as P.original. It will show the original speech.
If you do not use this trick, the speech might be without the punctuation symbols.
Comments
Post a Comment