Overview

CX AI Features

// Introduction

Today’s consumers have an abundance of choice in a borderless and boundless digital landscape. They’re increasingly impatient with brands that don’t deliver on their expectations. Every engagement is a piece of the experience, and builds on one another.

Information is power, and with CX AI powered by over 20 years of contact center research and innovation, ElevateAI enables immediate insight into interactions at scale including voice activity and purposeful, pre-built behavioral models.

We are pleased to announce that effective v1.11 on February 29, 2024 all declared interactions will include the available CX AI processing.

ElevateAI provides access to an AI framework that enables you to make every application and process smarter:

  • Complete, objective, and automated analysis of your interactions
  • Predictive and prescriptive insights
  • Purpose-built and ready made models


// Voice Activity

For any audio interaction, the voiceActivitySegments included when retrieving CX AI results provide the start and stop time in milliseconds for all voice and silence segments, enabling immediate insight into total non-interaction time (reflective of the time participants are not interacting with one another as a result of holds and/or silences).



In addition, for audio interactions, ElevateAI automatically identifies whether each voice segment is associated with particpantOne or particpantTwo, empowering you with information related to the total talk time associated with each speaker as well as any cross-talk that may have occurred and who initiated it.

Sample CX AI Response



// Enlighten Bundles

NICE Enlighten scores associated with each behavioral model are represented as an overall score as well as through each quartile of the interaction, providing key insight into shifts in behaviors.

Behaviors for Customer Satisfaction

// Acknowledge Loyalty

A more positive Acknowledge Loyalty score reflects higher predicted agent proficiency in taking a moment to acknowledge the caller's tenure with the organization, and showing appreciation for their loyalty. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.

// Active Listening

A more positive Active Listening score reflects higher predicted agent proficiency in in actively responding in the conversation, and not asking the caller to repeat themselves. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Be Empathetic

A more positive Be Empathetic score reflects higher predicted agent proficiency in acknowledging stated issues and their related impacts/hardships to the caller. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Build Rapport

A more positive Build Rapport score reflects higher predicted agent proficiency in engaging the caller in general dialogue not specific to the reason for contacting to build a personal connection. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Demonstrate Ownership

A more positive Demonstrate Ownership score reflects higher predicted agent proficiency in reassuring the caller that the agent understands the issue, and is ready and able to help. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Effective Questioning

A more positive Effective Questioning score reflects higher predicted agent proficiency in asking meaningful questions to explore the caller's experience, issues, and/or opportunities. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Inappropriate Action

A more positive Inappropriate Action score reflects higher predicted agent proficiency in not denying a caller's request to transfer the contact, using inappropriate language, or other offensive acts. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Promote Self Service

A more positive Promote Self Service score reflects higher predicted agent proficiency in promoting the availability of self-service options (IVR, website, app, etc.) where appropriate during an interaction. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



// Sentiment

A more positive Sentiment score reflects higher predicted likelihood the customer would provide a positive post-contact survey upon completion of the interaction. Detectors include both language properties (spoken words) and acoustic properties (pitch and tone, laughter detection, and cross-talk). The model was trained to consider language across both participants when identifying the caller's sentiment, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

// Set Expectations

A more positive Set Expectations score reflects higher predicted agent proficiency in summarizing actions and next steps and informing the caller of what to expect and/or required actions. The model was trained to consider language across both participants when identifying the agent's demonstrated proficiency, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this agent behavior.



Vulnerable Customer Behaviors

A vulnerable customer is defined by the UK Financial Conduct Authority as "someone who, due to their personal circumstances, is especially susceptible to harm, particularly when a firm is not acting with appropriate levels of care." These customers may require additional actions on the part of financial firms to ensure they receive fair and equal service.

// Vulnerable Customer

A more positive Vulnerable Customer score reflects higher predicted probability that the customer falls into one or more of the five Vulnerable Customer sub-categories. The model was trained to consider language across both participants when identifying the customer's demonstrated vulnerability, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

// Vulnerable Capability

A more positive Vulnerable Capability score reflects higher predicted probability that the customer has reduced knowledge of financial matters or low confidence in managing money; and/or a low capability in other relevant areas such as literacy or digital skills. The model was trained to consider language across both participants when identifying the customer's demonstrated vulnerability, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

// Vulnerable Health

A more positive Vulnerable Health score reflects higher predicted probability that the customer is experiencing physical health conditions or illnesses that affect his/her ability to carry out day-to-day tasks. The model was trained to consider language across both participants when identifying the customer's demonstrated vulnerability, and as such it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

// Vulnerable Life Event

A more positive Vulnerable Life Event score reflects higher predicted probability that the customer is experiencing negative circumstances such as bereavement, job loss, or relationship breakdown. The model was trained to consider language across both participants when identifying the customer's demonstrated vulnerability, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

// Vulnerable Mental Health

A more positive Vulnerable Mental Health score reflects higher predicted probability that the customer is experiencing mental health conditions or illnesses that affect his/her ability to carry out day-to-day tasks. The model was trained to consider language across both participants when identifying the customer's demonstrated vulnerability, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

// Vulnerable Resilience

A more positive Vulnerable Resilience score reflects higher predicted probability that the customer is experiencing reduced ability to withstand financial and/or emotional shocks. The model was trained to consider language across both participants when identifying the customer's demonstrated vulnerability, and as such, it is recommended that the score associated with allParticipants is leveraged when assessing this customer behavior.

Financial Distress (BETA)

High Risk Transaction (BETA)