Table of Contents

Kyubo

This table displays a detailed summary of customer service sessions, allowing evaluation of compliance with configured service thresholds, identification of attention delays, and analysis of agent performance.

General structure
Each row represents a unique interaction session between a customer and the system (either bot or human agent), and includes information such as:

  • Session ID and channel
  • Date and time the agent took the interaction
  • Associated workgroup
  • Measured values and defined thresholds for various service levels
  • Response and abandonment times
  • Agent identity (if applicable)

Data Interpretation

The table includes the following columns:

  • SessionId: Unique identifier of the session.

  • Channel: Channel from which the interaction originated.

  • Agent connection date: Date and time when the agent responded to the interaction.

  • WorkGroup: Workgroup or use case to which the session belongs.

  • Session service level [sec]: Time from when the customer initiates the session until an agent responds to the interaction.

    Note

    The agent is considered to have "responded" the moment they click on the customer's interaction in the panel, marking the start of active attention by the agent.

  • Derivation service level [sec]: Time from the customer's request for transfer until the agent responds to the interaction.

  • First response service level [sec]: Time between the customer’s transfer request and the agent’s first reply.

  • Interaction service level [sec]: Time elapsed from the derivation request to the agent’s first message.

  • Compliance: Percentage of compliance with service thresholds.

  • ART (Accumulated Response Time): Total maximum response time in hh:mm:ss format.

  • DPC [sec]: Difference between planned and actual agent connection time (Delay Planned vs Connected).

  • Abandonment source: Who abandoned the session (customer or agent).

  • Agent: Name of the agent who handled the session.

To properly visualize service levels, it is necessary to establish the thresholds against which the recorded times will be compared. These thresholds are configured in the Derivation to Kyubo action.

If the measured time is below the threshold, the value will appear in green, indicating it is within the acceptable range. If it exceeds the threshold, it will appear in red, indicating a time exceedance alert.

service-levels

Service Level Defines the time from... To...
Session Customer session start Agent responds
Derivation Customer requests transfer Agent responds
First response Customer requests transfer Agent's first message
Interaction Customer replies to agent Agent's next response

If you want to express the Interaction service level as a formula, it could be written as:

AgentWaitTime = Timestamp(Agent_n) - Timestamp(FirstCustomerMessageAfterAgent_n-1)

Where:

  • Agent_n: current agent message.
  • Agent_n-1: previous agent message.
  • FirstCustomerMessageAfterAgent_n-1: first customer message sent after Agent_n-1.

Use Case

Real-time monitoring and improvement of operational performance

A customer service supervisor uses this table to monitor active and recent sessions, quickly identifying whether agents are responding within the SLA (Service Level Agreement) timeframes. If the values of "Session service level", "Derivation", or "First response" exceed the defined thresholds, immediate actions can be taken, such as:

  • Assigning more agents to the affected queue.
  • Prioritizing urgent transfers.
  • Contacting the agent with the most delays to evaluate the cause.

Additionally, analyzing fields like DPC, ART, and Compliance helps detect inefficiency patterns, plan better shift distributions, and evaluate individual agent performance. These insights are key to continuous feedback processes, training, and operational improvement.