Acoustic Sentiment Analysis for Dynamic Call Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional customer service call routing systems lack the ability to dynamically allocate calls based on customer emotions and agent sentiment handling capabilities, leading to inefficient matching of customers with suitable agents.
Innovation Solution
A method and system that determine sentiment indicators from acoustic parameters of customer calls, select candidate agents based on their sentiment handling capabilities, and allocate calls dynamically, using historical data and real-time analysis to ensure effective emotion handling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional call-assignment processes such as round-robin or skills-based routing are used, then calls are routed based on basic agent skills and queue management, but the system cannot analyze customer emotions or match agents to calls based on sentiment handling capabilities
Solution Approach 1:
The patent introduces an acoustic analysis engine as an intermediary component that processes customer voice calls and extracts sentiment indicators. This intermediary translates raw voice data into structured emotional profiles that the routing engine can use for intelligent call allocation, bridging the gap between traditional routing systems and emotion-based matching without requiring complete system replacement
Solution Approach 2:
The patent replaces traditional mechanical routing methods (round-robin, basic skills-based routing) with an intelligent system that uses acoustic parameter analysis and machine learning algorithms. The routing engine substitutes deterministic allocation rules with adaptive decision-making based on real-time emotional analysis, enabling dynamic call distribution that responds to customer sentiment states
2Productivity
If agents are assigned to only one queue of incoming calls sequentially, then agents can focus on specific call types, but agents who can deal with a range of call types cannot effectively utilize their diverse skills
Solution Approach 1:
The patent implements dynamic call allocation where agent assignments are not fixed but continuously adjusted based on real-time acoustic analysis of customer emotions. The system adapts routing decisions moment-to-moment, matching agents to calls based on current sentiment indicators and historical performance data, enabling flexible utilization of multi-skilled agents across different call types while optimizing for emotional compatibility
Solution Approach 2:
The patent changes the routing parameters from static agent skills and queue assignments to dynamic acoustic parameters including speaking intensity, speaking rate, pitch variation, and other voice-based emotional indicators. By monitoring these parameters in real-time and adjusting routing decisions accordingly, the system enables agents to handle diverse call types based on their demonstrated ability to manage specific emotional states rather than fixed skill categories
3Measurement precision
If skills-based routing assesses skills by dialed number and IVR choices, then routing decisions are made quickly, but the system cannot measure customer emotions or agent ability to handle emotions
Solution Approach 1:
The patent performs preliminary acoustic analysis of customer voice calls during the initial interaction phase, extracting sentiment indicators before final routing decisions are made. By analyzing acoustic parameters early in the call flow and pre-computing emotional profiles, the system prepares matching data in advance, enabling rapid routing decisions that incorporate detailed emotion detection without significant time delays
Solution Approach 2:
The patent implements continuous acoustic monitoring throughout the call duration, not just at initial routing. The system continuously analyzes speech patterns, tone, and emotional indicators, maintaining an updated emotional profile that can trigger dynamic routing adjustments during the conversation. This continuous action ensures accurate emotion detection while distributing processing load over time, preventing bottlenecks in call allocation
4Adaptability or versatility
If traditional routing systems allocate calls without analyzing historical conversations, then allocation is simple and fast, but the system cannot apply knowledge from past interactions to improve future routing
Solution Approach 1:
The patent implements feedback loops where outcomes of previous call allocations are analyzed to improve future routing decisions. The system tracks which agents successfully handled calls with specific emotional profiles and uses this historical performance data to refine matching algorithms. Acoustic parameters from past conversations are stored and reused, allowing the system to learn from historical interactions and continuously improve routing accuracy without requiring complex real-time analysis of every past call
Data Source
AI summary
The present disclosure relates to methods of systems for allocating a call from a user to an agent. Embodiments of the disclosure may determine a set of sentiment indicators associated with the user from one or more acoustic parameters of the call. In addition, embodiments of the disclosure may select a candidate agent to handle the call based on the set of sentiment indicators and a sentiment handling capability associated with the candidate agent. Moreover, embodiments of the disclosure may allocate the call to the candidate agent.