Symptom / Need

I need to get detailed analysis of agent-customer interactions, including silent times, greetings, customer sentiment, and keyword detection in the conversation.

Context / Scenarios

Businesses looking to improve service quality, evaluate agent communication, and analyze customer sentiment require accurate information about what happens during interactions. wolkvox's speech analytics system allows these aspects to be identified by differentiating conversation participants and detecting key patterns in calls and chats.

Answer (Solution/Recommendation)

To obtain a detailed analysis of interactions using speech analytics, you can generate the "1. CDR Speech" report in the "Speech Analytics" tab within the "Reports" section of wolkvox Manager.

This report includes the following key columns:

• AGENT_ID: Extension number of the agent who attended the interaction.
• AGENT_NAME: Name of the agent who handled the call or chat.
• AGENT_DNI: Agent identification number.
• CONN_ID: Unique identification number of the call or interaction.
• TIME: Total duration (in seconds) of the interaction.
• DATE: Date the call or chat occurred.
• SILENCE: Number of seconds of silence detected in the conversation. When the conversation channel is other than “voice” (call), this value will always be 0.
• CUSTOMER_PHONE: Customer phone number.
• COD_ACT: Activity code or typing assigned to classify the result of the call or chat.
• DESCRIPTION_COD_ACT: Description of the activity code used in the interaction.
• GREETING: Indicates whether the agent greeted you at the beginning of the conversation and what words they used to do so. Its format is:
  • "yes - [greeting phrase]" (if the agent greeted).
  • "no" (if the agent did not greet).
• SLOW_TO_GREET: Indicates whether the agent took a long time to greet:
  • "yes" (if he was slow in greeting).
  • "no" (if he greeted in time).
• GREETING_AGENT_NAME: Indicates whether the agent mentioned your name when greeting. Its format is:
  • "yes - [name said]" (if the agent mentioned your name).
  • "no" (if he didn't mention it).
• AUDIO_TEXT: Converts the audio conversation to text, allowing you to view the spoken content in the interaction.
• CUSTOMER_ID: Customer identification number.
• FEELING: Sentiment detected in the customer's voice during the conversation. Possible values are:
  • "very negative" (very negative).
  • "negative" (negative).
  • "neutral" (neutral).
  • "positive" (positive).
  • "very positive"
• CHANNEL: Channel used for interaction. It can contain the following values:
  • "Facebook chat"
  • "instagram chat"
  • "chat-sms"
  • "web chat"
  • "whatsapp chat"
  • "voice" (call).
• Custom Categories: These are up to 10 configurable columns in the “Voice and Text Analytics” section of wolkvox Manager (in the left side menu). These custom columns go after the “GREETING_AGENT_NAME” column until just before “AUDIO_TEXT”. These categories allow you to identify specific keywords or phrases within the conversation to assess the quality of service.