Voice Agent Function Extraction from Usage Logs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice agent development processes, such as those for Samsung Bixby, Amazon Alexa, and Google Assistant, often miss important user requests due to a top-down approach that considers user feedback only after functions are published, leading to unsupported intentions and costly market research methods that are limited in capturing user demands.
Innovation Solution
A method and apparatus that analyze usage logs to extract user utterances, generate proto capsules, determine importance ranks, and create source code stubs for new functions based on high-ranked proto capsules, enabling direct and accurate response to user requests without the need for extensive market research.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a top-down approach is used to develop voice agent functions based on business demands, then functions can be published systematically, but user requests beyond current functions are classified as unsupported and important user demands are missed
Solution Approach 1:
The system performs preliminary analysis of usage logs to identify unsupported functions before they become significant user complaints. By proactively analyzing user attempts at unsupported functions and generating candidate functions in advance, the system prevents the accumulation of unmet user needs while maintaining systematic development processes.
Solution Approach 2:
The system establishes a continuous feedback loop where usage logs are automatically analyzed to identify patterns of unsupported function requests. This feedback mechanism transforms previously lost user voice into actionable insights, enabling the development team to prioritize and implement new functions based on actual user demand rather than assumptions.
2Loss of information
If market research methods such as surveys and interviews are used to capture user voice, then customer feedback can be collected, but the process is costly and limited in capturing comprehensive user demands
Solution Approach 1:
The system enables self-service analysis by automatically processing usage logs without requiring external market researchers. The automated analysis pipeline extracts user intents, clusters similar requests, and generates candidate functions autonomously, eliminating the need for costly manual surveys and interviews while comprehensively capturing all user interactions.
Solution Approach 2:
Instead of conducting expensive primary market research, the system copies and analyzes existing usage log data that already contains authentic user expressions. This approach leverages readily available data to understand user needs, avoiding the costs and limitations of traditional market research methods while maintaining comprehensive coverage of user demands.
3Measurement precision
If usage logs are analyzed manually to identify new functions, then user feedback can be understood, but the process is time-consuming and cannot keep pace with increasing number of functions and usage logs
Solution Approach 1:
The system replaces manual mechanical analysis with automated computational processing. Natural language processing algorithms automatically parse usage logs, identify user intents, and cluster similar requests, achieving both high precision in understanding user intent and rapid processing capability that scales with increasing data volume.
Solution Approach 2:
The system changes the parameter of analysis speed from manual to automated processing. By implementing automated text analysis pipelines with configurable parameters for intent recognition and function generation, the system can process vast amounts of usage log data quickly while maintaining or improving the precision of user intent understanding compared to manual analysis.
4Adaptability or versatility
If the range of functions of a voice agent increases, then users can explore more capabilities, but the number of usage logs increases and unsupported functions become more frequent
Solution Approach 1:
The system performs preliminary analysis of usage logs to identify patterns of unsupported function requests before they accumulate. By proactively detecting and addressing gaps in function coverage through automated analysis, the system maintains high reliability even as the function range expands, preventing the frequency of unsupported responses from increasing.
Solution Approach 2:
The system establishes a continuous feedback mechanism that monitors usage logs to detect when users attempt unsupported functions. This real-time feedback enables the system to identify and prioritize new function development needs, ensuring that the expanding function range is matched by corresponding improvements in coverage, thereby maintaining reliability.
Data Source
AI summary
A method and apparatus for generating a new function of a voice agent, wherein usage logs of users of the voice agent may be analyzed to extract a set of utterances of the users with respect to a new function of the voice agent, proto capsules for the set of utterances are provided. The method includes based on the set of utterances, ranks of importance of the proto capsules may be determined, a vocabulary of a proto capsule having a higher rank than a preset criterion may be identified, and a source code stub for a new function of the voice agent corresponding to the proto capsule having the higher rank may be generated based on the identified vocabulary.