Amazon Lex is a service for building conversational interfaces into any application using voice and text. With Amazon Lex, you can quickly and easily build conversational bots (chatbots), virtual agents, and interactive voice response (IVR) systems. By leveraging foundation models (FMs) to enhance its natural language understanding capabilities, Lex can now resolve slot values in user utterances with greater accuracy. The Assisted Slot Resolution feature uses the advanced reasoning capabilities of FMs to improve accuracy and ultimately a better customer experience.
Assisted Slot Resolution is only triggered upon a fallback. For example, if an IVR asks a user “How many people would you like to include on the reservation?” and the user responds “I’d like to make a reservation for my husband, two kids, and myself,” a traditional bot might have difficulty resolving this to a numerical value. When this feature is enabled, upon slot retry, Lex will use an LLM to assist with understanding the user response by providing the correct value of 4. This feature is available for the following slot types: AMAZON.City, AMAZON.Country, AMAZON.Number, AMAZON.Date, AMAZON.AlphaNumeric (without regex), AMAZON.PhoneNumber, and AMAZON.Confirmation. This feature is available in IAD and PDX regions for Amazon Lex V2. Assisted Slot Resolution is only available for English speaking locales to start. To learn more, please see the documentation on Assisted Slot Resolution.