Physical Store Virtual Assistant with LLM-Generated Video Responses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Providing quick and accurate responses to customer questions and comments in a physical store environment is challenging, requiring a broad set of skills that include completing sales, showing alternative products, and understanding the subtleties of conversations, which can be difficult even for professional sales and support staff.
Innovation Solution
An interactive video interface (IVF) in a physical store equipped with a video camera, interactive screen, and synthetic human that uses large language models (LLMs) to collect user input, generate responses, and produce video segments for engaging customer interactions, including skills like product recommendations and ecommerce purchases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If professional sales and support staff are used to handle customer inquiries, then customer service quality can be maintained, but response speed and accuracy deteriorate due to the broad skill set required and time constraints
Solution Approach 1:
The system enables self-service through the synthetic human assistant that autonomously handles customer inquiries without requiring human sales staff intervention. The synthetic human collects user input, generates responses using LLMs, and performs skills like product recommendations and ecommerce purchases independently, freeing human staff from routine tasks while maintaining service quality and increasing response speed.
Solution Approach 2:
The patent replaces the mechanical system of human sales staff with an automated synthetic human assistant powered by large language models. This substitution eliminates the limitations of human response time and skill breadth while maintaining or improving service quality through consistent, accurate, and rapid responses to customer inquiries.
2Adaptability or versatility
If human sales staff are trained to handle diverse customer needs, then service comprehensiveness improves, but training time and cost increase
Solution Approach 1:
The synthetic human assistant is designed with universal capabilities to handle diverse customer needs through multiple skills including product recommendations, ecommerce purchases, and coordinating deliveries. The large language model provides the synthetic human with broad knowledge across product categories and customer service scenarios, eliminating the need for extensive human training while achieving comprehensive service coverage.
Solution Approach 2:
The system performs preliminary action by pre-loading the large language model with extensive product knowledge and service protocols before deployment. This allows the synthetic human to immediately handle diverse customer inquiries without requiring ongoing training, as the knowledge base is already established and ready for rapid application to various customer needs.
3Reliability
If more skilled sales personnel are hired to handle complex inquiries, then customer service capability improves, but operational cost increases
Solution Approach 1:
The patent creates a synthetic human copy that replicates and enhances the capabilities of skilled sales personnel. Instead of hiring multiple highly-trained human staff members, the system uses a single synthetic human powered by large language models to provide equivalent or superior service capability at lower operational cost, as the synthetic human can handle multiple customers simultaneously without additional per-personnel costs.
Data Source
AI summary
Techniques for video processing using artificial intelligence are disclosed. An interactive video interface (IVF) located in a physical store is accessed. The IVF includes a video camera, interactive screen, and one or more microphones. The IVF includes a synthetic human that interacts with customers. A customer initiates an interaction with the IVF by standing within a minimum distance of the IVF, speaking to the IVF, interacting with the IVF screen, or scanning an identification card. The IVF collects input from the customer regarding products in the store. The IVF creates responses to the customer inquiries based on a large language model (LLM), including skills such as checking product inventory, recommending related products, completing ecommerce purchases, and coordinating product deliveries. The IVF produces a video segment featuring the response performed by the synthetic human. The IVF presents the video segment to the customer, including performing the selected skills.