Slot filling extracts predefined conversational parameters from a user’s utterance so a dialog system can execute the requested action. Think of each slot as a required function argument: “What’s the weather in Paris?” supplies a weather intent plus a location, while “Will it rain next Sunday?” supplies timing. section 12.2.1
Slots differ from entities. An entity is information a bot recognizes, whether or not it matters to the request; a slot is explicitly predefined in the interaction model as information the bot must look for. Common slots include locations, times, durations, and distances, while domain-specific systems may define drug names or crop types. section 12.2.1
Weather request
For “What’s the weather in Paris?”, the system identifies the weather intent and fills the location slot with “Paris.” For “Is it going to rain next Sunday?”, it fills the timing slot with “next Sunday.” section 12.2.1
