Speech analytics is a powerful tool that transforms recorded conversations into actionable data. By analyzing speech patterns, identifying key phrases, and categorizing conversations, businesses can gain valuable insights into customer interactions, improve agent performance and business KPIs, and enhance overall customer experience. This guide delves into the foundational concepts of Tags and Categories in speech analytics, explaining their roles, functionalities, and applications.
Understanding and effectively implementing the concepts of Tags and Categories is essential for maximizing the potential of speech analytics. By accurately capturing and categorizing key elements of conversations, businesses can derive meaningful insights, improve customer service, and drive strategic improvements. You can use this guide as a foundation to build a robust speech analytics framework tailored to your unique business needs.
Tags Concept
Definition
Tags are the most granular elements in speech analytics. They consist of specific words, phrases, or patterns identified within conversations. Tags are building blocks for more complex analytical constructs and are crucial for detailed analysis and reporting.
Tags Applications
Filters: In searching through large volumes of recorded conversations, tags act as filters to quickly locate interactions where specific terms were mentioned, enhancing the efficiency of data retrieval.
Analytics: Tags provide a foundation for deeper analysis by highlighting specific conversation terms. For instance, tagging words related to dissatisfaction can help identify problematic interactions.
Reports: Tags contribute to comprehensive reports by aggregating data on how often certain phrases were used by agents or clients, helping pinpoint trends and improvement areas.
Tags Functionality
Word and Phrase Matching: Tags are configured to recognize specific words or phrases within conversations. This can be tailored for both agent and client channels, enabling precise identification of the presence or absence of dialogue elements.
Customization: Tags can be customized to match any relevant term or phrase, making them highly versatile for various business needs. Tags can consist of single words (e.g., 'dissatisfied,' etc.) or include whole phrases (e.g., 'want to find')
Real-time and Post-call Analysis: Tags can be applied in real-time to flag conversations as they occur or analyzed post-call to extract data retrospectively.
Examples
Consider a customer service center where common customer concerns include billing issues and technical support. Tags can be created for phrases like "billing problem," "invoice error," "technical issue," and "software bug." These tags enable quick identification of conversations involving these issues, aiding in faster resolution and targeted training.
Categories (Topics) Concept
Topic classification for post-conversation analytics helps identify and categorize the key issues or themes discussed during customer interactions. This allows contact centers to gain insights into common problems, identify trends and products, and optimize agent training and business processes.
In Ender Turing, the Topic is called Category to highlight that it can be used for much broader purposes than just a topic of conversation. For example, If you need to categorize Product-123 related conversation and analytics, you can use this categorization.
Categorization Approaches
This guide will walk you through creating topics in the Ender Turing application, covering three distinct categorization approaches that can be mixed depending on your tasks and goals:
Generative Categorization (using Generative AI): This option leverages AI models to automatically generate and categorize topics based on the call's content. The AI identifies and categorizes key themes without predefined rules, adapting to new topics over time. It’s highly flexible and works well in dynamic environments where new trends emerge frequently. Generative categorization is a fast and efficient way to categorize new or evolving topics but can't be used to categorize through historical conversation.
Rule-Based (based on Filters and Tags): This approach categorizes topics using predefined rules, such as specific keywords, tags, or filters. It’s ideal for contact centers with well-established categories where certain phrases or terms reliably indicate specific topics (e.g., “refund request” or “technical issue”). Rule-based categorization is a fast and efficient way to categorize new and historical topics but is less adaptable to new or evolving topics.
Manual (Human Categorization): In some cases, you need to manually link the Category to the conversation, either for further training or an internal process. Using manual category, the user can select an appropriate Category from a predefined list.
Categories Applications
Topic Analysis: Categories facilitate the analysis of overarching topics by aggregating data from related tags. For instance, a category like "Customer Complaints" might include tags related to various complaints, providing a comprehensive view or even more complex rules to classify and group types of complaints.
Performance Metrics: By categorizing interactions, businesses can measure the frequency and impact of specific topics and react to emerging patterns, aiding in performance metrics and strategic decision-making.
Training and Development: Categories help identify common themes in customer interactions, guiding the development of targeted training programs for agents.
Topics & Categories Functionality
Advanced Logic: Categories are built using multiple tags, creating a composite view of related terms and phrases. This helps recognize broader subjects discussed in conversations.
Topic/Subject Detection: By grouping tags into categories, the system can detect and analyze broader topics such as customer sentiment, product feedback, or service complaints.
Examples
In a retail customer service environment, categories can be created for broad topics like "Product Issues," "Shipping Problems," and "Customer Feedback." Each category would encompass various tags such as "defective product," "late delivery," "positive feedback," and "suggestions for improvement." This allows for a holistic view of customer interactions related to each major topic.
Best Practices
Regular Updates: Continuously update tags and categories to reflect changing business priorities and emerging trends in customer interactions.
Collaboration: Work closely with stakeholders across departments to ensure tags and categories align with organizational goals.
Training: Provide thorough training to staff on effectively utilizing tags and categories within the speech analytics platform.
Configure Tags and Categories
Usual Speech Analytics Configuration Use Cases
Impact inside of Contact |
|
Quality Assurance |
|
Improve Customer Satisfaction (Experience) | NPS/CSAT |
Follow Internal Processes (Data availability, CRM, Orders, etc.) | Operation Improvement |
Script Adherence |
|
Compliance |
|
Verification | Risk mitigation |
Stop words (regulatory requirements) | Risk mitigation |
List of Consents before selling product/service inside a call | Risk mitigation |
Sales in Contact Center |
|
Rejection Reasons Understanding for Further Objection Handling Scripts | Conversion Increase |
Proposal of Specific products or services in every conversation | Sales Increase |
Impact outside of the Contact Center |
|
Marketing |
|
Feedback on the Marketing List for Outbound calling |
|
Measurement of Marketing Campaign (Are customers aware? Do they ask about the program) |
|
Price changes - An increase in price leads to more rejection in conversation with Price reason |
|
Industry Specific |
|
Request for Specific Doctors (Medical) | Schedules Planning |
Request for Specific Products | Resource Planning |
Insurance Approvals for specific diseases or Medicaments | Cost Control |
Step-by-Step Process
Determine key business KPIs and processes to improve
Ideate Key Conversation Elements matching the above KPIs and processes: Use Ender Turing AI Library to generate.
Identify Key Phrases and Words: Determine the specific words and phrases relevant to your business needs. Use Ender Turing AI Library to generate.
Create Tags: Follow the guide to configure tags to match these phrases within your speech analytics tool.
Create Categories: Develop categories by grouping related tags under broader themes.
Analyze and Report: Use the tagged data and categorized topics to generate insights, produce reports, and inform business strategies.