U-Self Serve AI Model Overview
Uniphore's U-Self Serve is an Intelligent Virtual Assistant with which Enterprises can automate customer support conversations in text or voice. U-Self Serve can also be used in Visual IVR mode where customers on the phone will also use chat to provide input to the Virtual Assistant to take the conversation forward.
U-Self Serve configuration contains two parts. The AI model which has the Virtual Agent with Intents, Entities and Training data. The Intent flows which contain the Conversation flows for each Intent.
The U-Self Serve AI data model is structured as follows:
Virtual Agent (VA) - An X-Console Account can run and manage multiple VAs. Each VA is intended to respond to requests (or questions) on a single customer-focused topic. The same VA can be run simultaneously on multiple customer channels (Example: device-based chat, text or voice). Each VA can be assigned multiple languages.
Entity - A term or phrase used to define an Intent (Example:, City, Auto License, Travel Date, Device ID). Typically, in AI NLP, an Entity receives its value from a customer's spoken or written expression.
All important data that is needed to fulfill an Intent require a corresponding Entity. Each U-Self Serve VA can support an unlimited number of Entities, as well as a wide array of Entity types. To execute runtime activities in X-Console, an Entity must be mapped to a specific Variable.
U-Self Serve uses the following Entity classes:
System Entity describes a generic value that is maintained on the Uniphore Natural Language Understanding (NLU) Server.
Example:, Date, Phone Number, Email Address, Amount of Money
Custom Entity describes any conversation-specific value that is maintained in the U-Self Serve AI Model
Example:, ToCity, Customer_Level, Benefit_Type, BillingMethod.
Alphanumeric Entity describes any information that contains Alphanumeric value.
Example: Social Security Number, Insurance Policy Number, Driving License Number
Intent - The customer's desired goal for the self-service experience (Example:, Change an Address, Book a Hotel Room, Order a New Service). Typically, in AI natural language processing (NLP), the Intent is derived from a customer's spoken or written expression. Each U-Self Serve VA can support an unlimited number of Intents. To execute runtime activities in X-Console, each Intent can be mapped to a specific Flow and Step.
Slot - A value entered or spoken by the customer in response to a runtime Flow question. The value is associated with an Entity and assists the VA to determine precisely what the customer needs.
An Slot value may be required in order for the Flow or U-Self Serve operation to perform successfully. An Slot value can be used by a relevant X-Designer element, and also combined with a AI generated response.
Slots can be added in the Training Phrases for a U-Self Serve data model.
Training Phrases - All possible user utterances are added as Training Phrases. The success of a Virtual Assistant depends on how we maintain and train it periodically.