Avatar Customer Service Feedback Using Purchase Behavior Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to set optimal avatars for customer service that effectively cater to diverse customer characteristics, leading to inefficiencies in sales measures due to the vast number of parameter combinations in prompt settings.
Innovation Solution
An information processing device analyzes customer behavior and features through video data, calculates occurrence rates of purchase behaviors for different avatar styles, and updates prompts based on these rates to optimize avatar settings for improved sales effectiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If generative AI is used to generate avatar images by designating prompts with keywords and weights, then avatar images can be generated with specific features, but it becomes impractical to specify an optimal prompt due to the vast number of parameter combinations
Solution Approach 1:
The patent segments the avatar generation process into distinct components: (1) video acquisition and customer behavior analysis, (2) feature classification into predetermined categories, (3) occurrence rate calculation for each class, and (4) avatar image generation based on selected classes. This segmentation transforms the complex task of manually tuning numerous prompt parameters into an automated process that selects from predefined feature classes, thereby resolving the contradiction between avatar customization capability and prompt parameter complexity.
2Productivity
If multiple customer service styles are created for different customer features, then optimal avatars can be provided for various customers to promote sales, but the system complexity increases due to the need to manage and update styles for each feature class
Solution Approach 1:
The patent implements a feedback mechanism where customer behavior data collected from video analysis is used to calculate occurrence rates of purchase-related behaviors for each feature class. These occurrence rates then feed back into updating the customer service styles, creating a closed-loop system that automatically optimizes avatar selection based on actual customer responses. This feedback mechanism eliminates the need for manual management of multiple customer service styles, resolving the contradiction between sales promotion effectiveness and style management complexity.
Solution Approach 2:
The system performs self-service by automatically analyzing customer behavior, calculating occurrence rates, and updating customer service styles without requiring manual intervention. The avatar generation system uses the calculated occurrence rates to autonomously determine which customer service styles are most effective for different customer feature classes, thereby promoting sales while avoiding the complexity of manual style management.
3Measurement precision
If video analysis is performed to calculate occurrence rates of purchase-related behaviors for each feature class, then customer service styles can be updated based on actual behavior data, but the processing time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-defining feature classes and their corresponding customer service styles before actual customer behavior analysis begins. The system has predetermined categories ready for classification, which allows video analysis to focus on matching observed behaviors to these pre-established classes rather than creating classifications in real-time. This preliminary preparation reduces processing time while maintaining measurement precision, as the computational framework is already in place to handle the behavior data efficiently.
Data Source
AI summary
A non-transitory computer-readable recording medium storing an avatar generating program for causing a computer to execute process includes: acquiring a video obtained by imaging an area that includes a product shelf on which a product is arranged; specifying a behavior of a person and a feature regarding the person by analyzing the acquired video; classifying the specified feature based on a predetermined condition; acquiring a customer service style that is expressed by an avatar displayed in a visually recognizable manner for each person and that is for the product; calculating an occurrence rate of a behavior related to purchase of the product, for each classified feature, based on the specified behavior, for the customer service style; and updating the customer service style, for each classified feature, based on the occurrence rate.