Real-time Speech Snippet Delivery for Contact Center Agent Transfers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In contact centers, customers often become frustrated when they need to repeat information to a second agent after a call is transferred, as the agent is not immediately informed about the context and progress of the initial conversation.
Innovation Solution
A process that uses a speech analyzer to identify keywords during a call, creating snippets of the conversation which are then ordered and played to the second agent before the call is transferred, allowing the agent to understand the context and tone of the conversation without needing to listen to the entire call.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the entire call is transferred to the second agent without preprocessing, then the agent receives complete information, but the agent cannot quickly grasp the context and customer frustration increases
Solution Approach 1:
The call audio is segmented into multiple snippets based on detected keywords. Instead of transferring the entire call, the system identifies and extracts relevant segments that contain key information about the customer's issue, emotions, and context. This segmentation allows the second agent to quickly access only the necessary portions of the conversation rather than listening to the complete call.
Solution Approach 2:
The system performs preliminary analysis of the call during the first agent's conversation by continuously monitoring for keywords and preparing snippets in real-time. This preliminary action ensures that when the call is transferred, the snippets are already identified and ready for immediate playback to the second agent, eliminating any delay in information delivery.
2Loss of information
If keywords are configured and speech analysis is performed in real-time, then relevant snippets are identified for the second agent, but the system complexity increases
Solution Approach 1:
A speech analyzer component acts as an intermediary between the call audio and the snippet identification process. This intermediary automatically performs keyword detection and snippet generation based on configured keywords, reducing the manual effort required and standardizing the information extraction process. The speech analyzer mediates between the raw audio stream and the information needs of the second agent.
3Productivity
If snippets are presented to the second agent before call transfer, then handling efficiency improves, but the process complexity increases
Solution Approach 1:
The system merges multiple functions into the call transfer process: keyword detection, snippet identification, snippet ordering, and audio playback are all integrated into the single call transfer operation. This merging allows snippets to be automatically prepared and presented to the second agent as part of the standard transfer workflow, improving efficiency without requiring separate manual steps for each function.
Data Source
AI summary
A process for updating a second agent about a call in a contact center comprises receiving a call at the contact center and connecting the call to a device associated with a first agent of the contact center. A processor is used to configure a list of keywords to detect during the call, and when a key word is detected (e.g., using a speech analyzer), a snippet of the call based on the detected keywords is identified. The snippets are presented to a second agent through a device associated with the second agent. The call is then connected to the device associated with the second agent.