1. Introduction: What is CX Insights?
CX Insights is an advanced application within Alotech's AI Navigator module that leverages the power of artificial intelligence (AI) to analyze your customer interactions (calls and chats). This tool allows you to gain a deep understanding of your call center's performance, customer satisfaction, and the root causes of interactions. With CX Insights, you can increase operational efficiency, improve the customer experience, and make data-driven strategic decisions.
Advantages You Will Gain with CX Insights:
Higher Customer Satisfaction: Create a more positive customer experience by analyzing customer sentiment and proactively identifying issues.
Enhanced Operational Efficiency: Increase operational efficiency by identifying frequently recurring problems and bottlenecks in processes.
Data-Driven Decision-Making: Steer your resources correctly by making decisions based on real data, not guesswork.
Proactive Problem Solving: Prevent customer dissatisfaction by identifying trends and potential issues before they become major problems.
Improved Agent Performance: Provide more effective feedback and identify training needs by clearly seeing agent actions and customer requests.
2. Setup and Profile Management
To start using CX Insights, you first need to create an analysis profile specific to your operation.
2.1. Access
To access the feature, follow these steps in the Alotech main menu:
Click on the System Administrator menu.
From the opened menu, go to the AI Navigator option.
Click on the AI Analysis submenu.
On this screen, you can manage your existing AI profiles and create a new CX Insights profile.
2.2. Creating and Configuring a New Profile
Click the New button in the top right corner of the AI Analysis page.
In the "Profile Settings" window that opens, select CX Insights as the application and click the Continue button.
In the Settings tab, configure the following fields:
Status: Determines whether the profile will be active. It must be in the Active position for the analysis to begin.
Profile Name: Enter a descriptive name for your profile (e.g., "General Sales Analysis").
Channel to be Analyzed: Choose which interaction types will be analyzed.
All: Analyzes both calls and chats. (Default)
Call: Analyzes only voice calls.
Chat: Analyzes only written chats.
Analysis Scope: Determine the scope of interactions to be included in the analysis.
All: All interactions are included in the analysis. (Default)
Queues: Only interactions belonging to specific queues you select are analyzed.
Agents: Only interactions belonging to specific agents you select are analyzed.
Category Creation Method: Choose how categories will be determined.
AI-Based: Categories are automatically created and named by artificial intelligence.
User-Defined: Categories are manually created and managed by you. (See Section 3 for details.)
2.3. Detailing the Analysis Scope
If you have selected Queues or Agents as the "Analysis Scope," the "Define Scope" area will become active.
Click the + Add Parameter button to add a new row.
In the opened row, select the relevant Queue or Agent.
From the Analysis Language column, select the analysis language for this queue/agent. Language selection is mandatory for each parameter.
If you want to assign the same language to multiple parameters, you can select the relevant rows and use the Bulk Assign Language button.
Duration Filter (Optional): This allows you to exclude calls shorter than the specified number of seconds from the analysis. For example, you can enter "10" in this field to exclude short calls under 10 seconds.
After completing all configurations, click the Save button.
Important Note: For CX Insights analysis to work correctly for calls, your operation must be using stereo (separate channel) audio recording. Please contact your project manager for this setting.
3. CX Insights Reporting Interface
After configuring your profile, go to the Reports > CX Insights menu to view the analyzed data.
3.1. Filtering
You can refine your data using the filters at the top of the reporting page:
Date Range: Displays data for a specific time period.
Start Time: Filters the interaction's start time for the relevant data.
End Time: Filters the interaction's end time for the relevant data.
All Channels: Allows you to filter the report by All, Call, or Chat.
Queue: Filters by queues.
Main Category: Filters by the main categories assigned to interactions.
Sub-Category: Filters by the sub-categories assigned to interactions.
Custom Variable Name: Specifies the custom variable name defined in the interaction.
Custom Variable Value: Filters by the value of the custom variable defined in the interaction.
Call Type: Inbound or Outbound Call.
3.2. Executive Summary Area
At the very top of the page, there are cards that provide a general summary based on the filtered data. These cards include key metrics such as Analyzed Calls, Analyzed Chats, Solution Provided, and Decisive Positive Sentiment Score.
Resolution Rate: Resolution is determined by a comprehensive analysis of the interaction, assessing whether the agent provided any form of solution to the customer's request. Resolution is marked as "Yes" when a solution or necessary guidance is provided, regardless of its content. Regarding the nature of resolution, CXI also identifies instances where the agent correctly assesses the customer's need and directs them to the right department or channel, or when the agent explains that no solution is available, as still providing the necessary support. Therefore, these cases are consistently marked as "Yes" to acknowledge that the customer's needs were addressed.
Decisive Positive Sentiment Score: This score measures the percentage of positive sentiment among all interactions that express a clear opinion (either positive or negative). It is calculated using the formula:
(Positive / (Positive + Negative)). Neutral interactions are excluded, thus offering a focused view on customer satisfaction.
3.3. Interaction Categories Table
This table is the heart of the analysis results. It shows which categories and sub-categories the interactions are concentrated in.
Columns: Provides detailed information such as Category Name, Sub-Category Name, Total Interaction Count, Chat Count, Call Count, Resolution, and Sentiment Distribution.
Advanced Filters: By clicking the filter icon in the top right of the table, you can filter the data in more detail:
All Interactions: Display categories and sub-categories with an interaction rate greater or less than a certain percentage.
Resolution: Filter for categories that are equal to, above, or below a certain resolution rate.
Sentiment Distribution: Focus on categories with a sentiment distribution (e.g., Negative) above a certain percentage.
3.4. Data Visualizations
CX Insights offers various charts to help you interpret data more easily:
Category Cloud: Visually highlights the most frequently discussed sub-categories. The size of the word represents the frequency. You can use the Frequency Filter to display categories above a certain number of conversations (e.g., more than 100).
Sentiment Analysis Chart: Shows the overall sentiment distribution of interactions (positive, negative, neutral).
Solution Provided Chart: Shows the ratio of resolved versus unresolved interactions.
Date-Based Solution and Sentiment: Displays the trends of sentiment and resolution rates along with the interaction count over the selected time period.
4. Analysis Results in Call and Chat Details
CX Insights not only provides general reporting but also displays AI analysis results in the details of each individual interaction.
4.1. Advanced Call Listening Screen
In operations where CX Insights is active, the call listening screen (Advanced Call Records) is presented with an enriched interface:
Call Transcription: A full text transcript of the conversation is located just below the audio recording. Speakers (Customer/Agent) are indicated separately. You can search within the text.
AI Call Review Panel: On the right side of the screen, the AI analysis results specific to that call are displayed:
Category and Sub-Category: The topic the call was assigned to.
Sentiment: The overall emotional tone of the call.
Resolution: Whether the customer's request was resolved.
Summary: A few-sentence summary of the call generated by AI.
Customer Requests: The customer's requests are listed in bullet points.
Agent Actions: The actions taken by the agent during the call are listed in bullet points.
Thanks to this unified view, you can quickly understand the content of a call by reading the transcript and summary instead of listening to long recordings, thereby speeding up your quality evaluation processes.
4.2. Chat Details Page
When you enter the details of a chat record, the CX Insights Results panel appears on the right side of the conversation text in the Chat History tab. This panel, similar to the call details, presents AI analyses specific to that chat, such as Category, Sub Category, Sentiment, Summary, Agent Actions and Customer Requests.
5. Tips and Best Practices
Descriptive Names: Ensure your profile and category names are descriptive so that everyone can easily understand them.
Correct Language Selection: Make sure you select the correct analysis language on a per-queue or per-agent basis for analysis accuracy.
Narrow the Scope: To optimize analysis costs, include only the necessary queues or agents in the analysis.
Stereo Recording: Confirm that stereo (separate channel) audio recording, a core requirement for the call analysis feature, is active in your operation.
Save Time: Instead of reviewing long interactions from start to finish, save time by first reading the Summary, Customer Requests, and Agent Actions sections generated by the AI.