Abstract: Systems and methods for mining question-and-answer pairs from conversation data to update knowledge bases are provided. In particular, a computing system may obtain conversation data, extract a question-and-answer pair from the conversation data with meta data using a machine learning model, the question-and-answer pair including a question and an answer corresponding to the question, determine whether a knowledge base includes existing content similar to the question-and-answer pair, provide the question-and-answer pair on a graphical user interface with an indication of a presence of any existing content in the knowledge base that is similar to the question-and-answer pair, receive an input to integrate the question-and-answer pair with the knowledge base, and update the knowledge base to integrate the question-and-answer pair in accordance with the input.

Abstract: Systems and methods for generating a routing recommendation for an incoming interaction in a contact center are described. For example, the method includes receiving interaction data associated with the incoming interaction, determining candidate agents based on one or more constraints, obtaining agent data related to the candidate agents, determining, for each of the candidate agents, an expected performance score using a predictive routing model, the expected performance score indicating a predicted performance of the corresponding candidate agent relative to a predetermined performance metric for handling the incoming interaction, and generating a routing recommendation for the incoming interaction based on the expected performance scores of the candidate agents, the routing recommendation identifying one or more agents from the candidate agents predicted to achieve a predefined level of the predetermined performance metric.

Abstract: A method for processing data for training a predictive routing model. The method includes receiving interaction data from previous interactions that includes audio data capturing a conversation and transcript data of the conversation. The method continues by performing speech analytics by processing the audio data to determine scores for speech metrics that include a measure of how much the agent or customer speaks during the conversation. The method continues by performing sentiment analysis to determine scores associated with sentiment metrics, the sentiment metrics including a measure of a sentiment based on classifying utterances appearing in the transcript data as being positive or negative. The method continues by performing feature engineering to generate feature data and generating a training dataset therefrom. The method continues by applying a machine learning algorithm to the training dataset to train a predictive routing model.

Abstract: A method for processing data for training a predictive routing model. The method includes receiving interaction data from previous interactions that includes audio data capturing a conversation and transcript data of the conversation. The method continues by performing speech analytics by processing the audio data to determine scores for speech metrics that include a measure of how much the agent or customer speaks during the conversation. The method continues by performing sentiment analysis to determine scores associated with sentiment metrics, the sentiment metrics including a measure of a sentiment based on classifying utterances appearing in the transcript data as being positive or negative. The method continues by performing feature engineering to generate feature data and generating a training dataset therefrom. The method continues by applying a machine learning algorithm to the training dataset to train a predictive routing model.

Abstract: A method for mining conversation flows according to an embodiment includes receiving a plurality of transcripts of conversations between contact center agents and users, generating a summary of each transcript of the plurality of transcripts by extracting, for each transcript, one or more intents associated with the respective transcript and one or more slot entries associated with the respective transcript, clustering the plurality of transcripts into a plurality of intent categories based on the respective summary of each transcript, wherein each intent category includes intents that are similar to one another, and analyzing each transcript within a selected intent category of to generate a guided flow for the selected intent category, wherein the guided flow defines a set of actions to be taken by a human/virtual contact center agent to resolve an intent associated with the selected intent category.

Abstract: A method for adaptive predictive routing in a contact center system according to an embodiment includes identifying an interaction to be routed to a contact center agent, determining, for each agent cohort of a plurality of agent cohorts in sequential order and for a cohort time period associated with the respective agent cohort, whether a contact center agent within the respective cohort is available to be routed the interaction, wherein the plurality of agent cohorts is in sequential order based on descending agent performance scores for at least one key performance indicator, and routing the interaction to a first contact center agent determined to be available to be routed the interaction.

Abstract: A method for adaptive predictive routing in a contact center system according to an embodiment includes identifying an interaction to be routed to a contact center agent, determining, for each agent cohort of a plurality of agent cohorts in sequential order and for a cohort time period associated with the respective agent cohort, whether a contact center agent within the respective cohort is available to be routed the interaction, wherein the plurality of agent cohorts is in sequential order based on descending agent performance scores for at least one key performance indicator, and routing the interaction to a first contact center agent determined to be available to be routed the interaction.

Abstract: A method for adaptive predictive routing in a contact center system according to an embodiment includes identifying an interaction to be routed to a contact center agent, determining, for each agent cohort of a plurality of agent cohorts in sequential order and for a cohort time period associated with the respective agent cohort, whether a contact center agent within the respective cohort is available to be routed the interaction, wherein the plurality of agent cohorts is in sequential order based on descending agent performance scores for at least one key performance indicator, and routing the interaction to a first contact center agent determined to be available to be routed the interaction.

Abstract: A method of routing interactions to contact center agents according to an embodiment includes identifying an interaction to be routed to a contact center agent, identifying a group of contact center agents as candidates for routing of the interaction, retrieving agent performance data for each candidate agent of the group of contact center agents identified as candidates for routing of the interaction, determining a predicted score for a key performance indicator for each candidate agent based on the agent performance data, determining an occupancy rate of each candidate agent based on the agent performance data, generating a ranking of the candidate agents for routing prioritization based on the predicted score for the key performance indicator for each candidate agent and the occupancy rate of each candidate agent, and signaling a routing device to route the interaction to a selected candidate agent based on the ranking of the candidate agents.

Abstract: A system and method are presented for customer journey event representation learning and outcome prediction using neural sequence models. A plurality of events are input into a module where each event has a schema comprising characteristics of the events and their modalities (web clicks, calls, emails, chats, etc.). The events of different modalities can be captured using different schemas and therefore embodiments described herein are schema-agnostic. Each event is represented as a vector of some number of numbers by the module with a plurality of vectors being generated in total for each customer visit. The vectors are then used in sequence learning to predict real-time next best actions or outcome probabilities in a customer journey using machine learning algorithms such as recurrent neural networks.

Abstract: A method of routing interactions to contact center agents according to an embodiment includes identifying an interaction to be routed to a contact center agent, identifying a group of contact center agents as candidates for routing of the interaction, retrieving agent performance data for each candidate agent of the group of contact center agents identified as candidates for routing of the interaction, determining a predicted score for a key performance indicator for each candidate agent based on the agent performance data, determining an occupancy rate of each candidate agent based on the agent performance data, generating a ranking of the candidate agents for routing prioritization based on the predicted score for the key performance indicator for each candidate agent and the occupancy rate of each candidate agent, and signaling a routing device to route the interaction to a selected candidate agent based on the ranking of the candidate agents.

Abstract: A system and method are presented for customer journey event representation learning and outcome prediction using neural sequence models. A plurality of events are input into a module where each event has a schema comprising characteristics of the events and their modalities (web clicks, calls, emails, chats, etc.). The events of different modalities can be captured using different schemas and therefore embodiments described herein are schema-agnostic. Each event is represented as a vector of some number of numbers by the module with a plurality of vectors being generated in total for each customer visit. The vectors are then used in sequence learning to predict real-time next best actions or outcome probabilities in a customer journey using machine learning algorithms such as recurrent neural networks.