1
votes

I have created a LUIS application on the portal, with several intents, trained and published.

https://docs.microsoft.com/en-us/azure/cognitive-services/luis/luis-quickstart-primary-and-secondary-data

When I make a GET request to the end-point I receive something like this:

{
  "query": "what is the weather",
  "topScoringIntent": {
    "intent": "Weather",
    "score": 0.972772241
  },
  "entities": []
}

On the face of it LUIS is only able to return me the best matching intent from my query.

Is there anyway achieving required slot filling, within the portal so that I could have an utterance like "What is the weather in City" and the response would come back with something suggesting there is a match but certain data is missing (the desired city to check) so I could inform the user what else they must provide before I generate an answer.

I intend on using LUIS in a class library which will return dialog for use with a telephony solution, if this helps. So no chat bot.

1

1 Answers

3
votes

This used to exist before LUIS way back when in Preview mode but did not make it into the GA release. Right now, this needs to be done with code. The BotBuilder-CognitiveServices library has some code you may be interested in using. There was also a blog post about it here.