Implementing FAQ’s is probably the most popular use case for a chatbot. ODA supports this with answer intents which allow to very easily being those FAQ’s under the same NLP ‘umbrella’ as the regular intents.
That being said, in some cases, where those FAQ’s where already present in a document (pdf) or in a webpage, porting that information to the ODA answer intents would take some manual work. Fear no more, because starting with 22.04 we can now import PDF’s and URL’s that will be converted into answer intents.
Read all about it here!
Create Skill
Let’s create a new Skill, and navigate to the new menu option called Knowledge.
The next step is to define exactly what kind of document we want to add.
Add Document
For this particular case I will add an FAQ for Star Wars, available as a webpage (imdb.com)
The selected options are URL and Language English.
Very simple actually. Once the button Create is pressed the process will take some minutes.
Once it completes we can Review Intents. This will allow us to modify the generated utterances, the answer and the questions as well. In this particular case I accept everything as is.
This will create answer intents automatically, with several utterances (variations from the original question) and the answer. All automated, which makes life really easy.
After training the model, everything is ready and works as any other answer intent!
Release 22.04 is rich in new cool features – including the new Visual Flow Designer! Stay tuned!