AI Support System Upgrade: Smarter Guidance for Doll Selection and Customer Assistance
At BreathyDoll, we continuously refine how users explore, understand, and choose companion dolls. With an expanding product range and increasingly detailed customization options, traditional static FAQs and manual support alone are no longer sufficient.
To improve clarity, accuracy, and decision efficiency, we have introduced a new AI-powered support system designed to assist users throughout the entire selection journey.
This system consists of two core components:
- AI Assistant (Knowledge-Based Support System)
- Doll Selection Assistant (Structured Recommendation Tool)
Together, they form an integrated decision-support framework that helps users make more informed and confident choices.
1. AI Assistant: Knowledge-Based Guidance Beyond Traditional Customer Support
The BreathyDoll AI Assistant is built on advanced large language model technology and operates on a fundamentally different principle compared to conventional customer service systems.
Instead of relying on simple keyword matching or static response templates, the AI Assistant:
- Integrates and synthesizes structured internal product knowledge
- Draws from accumulated industry expertise and long-term usage insights
- Generates responses based on contextual understanding rather than fixed scripts
- Provides consistent, grounded, and structured guidance across all product categories
A key design principle is knowledge-grounded response behavior, meaning the assistant does not generate speculative answers outside of its knowledge base. This ensures informational consistency and reduces the risk of misleading recommendations.
From a user perspective, the system functions as:
- A 24/7 available consultation interface
- A consolidated layer of product and material expertise
- A structured advisor capable of handling complex, multi-factor questions
In practice, it behaves less like a chatbot and more like a collective expert system built on accumulated domain knowledge.
2. Doll Selection Assistant: A Structured Decision System for Product Configuration
Selecting a companion doll involves balancing multiple constraints across materials, structure, durability, realism, and customization options.
To simplify this process, we developed the Doll Selection Assistant, a structured recommendation system that translates user preferences into ranked configuration plans.
2.1 How the System Works
The assistant is built around eight core decision factors, each representing a key concern in real-world selection behavior.
These factors are assigned different priority weights based on:
- material performance constraints
- structural compatibility
- durability considerations
- real usage scenarios
Instead of treating all preferences equally, the system applies a priority-weighted scoring model to ensure that physically constrained attributes are properly handled.
This prevents unrealistic combinations and ensures every recommendation remains technically feasible.
2.2 Why Priority Matters
Different materials and structural systems inherently involve trade-offs:
- TPE offers higher flexibility and generally performs better under frequent use scenarios
- Silicone provides more refined surface detail and higher visual realism
- Structural design choices such as head type determine compatibility with specific functional requirements
For example:
- If a user prioritizes frequent usage, durability constraints may shift recommendations toward TPE-based configurations
- If a user prioritizes high realism, silicone-based options may receive higher weighting due to surface detail advantages
2.3 Structural Constraint Example
Certain customization features are directly dependent on head and material architecture:
- If a user requires oral functionality (insertable mouth design), a soft head system is required, while hard head options are excluded due to structural limitations
- Hair implantation stability is generally higher on hard head systems
- On soft head systems, implanted hair may have reduced long-term stability and higher risk of detachment under extended use
These constraints are not arbitrary preferences but structural compatibility rules embedded into the recommendation logic.
2.4 Output: Three Ranked Configuration Plans
Based on user inputs, the system generates three complete configuration options, ranked by total score.
Each recommendation includes:
- full body material selection
- head system configuration
- compatibility-validated feature set
- optional upgrade components
- explanation of why each option is selected or excluded
This ensures that users not only receive a recommendation, but also understand the reasoning behind each configuration decision.
2.5 What Makes This System Different
Unlike traditional filters or static product listings, the Doll Selection Assistant:
- incorporates real manufacturing and structural constraints
- resolves conflicting user preferences through weighted prioritization
- outputs complete, actionable configuration plans
- provides transparent reasoning behind recommendations
Even first-time users can clearly understand how materials, structure, and customization interact in real product design.
3. A Unified Support Experience
Together, the AI Assistant and Doll Selection Assistant form a unified support ecosystem:
- The AI Assistant provides contextual explanations and knowledge-based support
- The Selection Assistant converts user preferences into structured product configurations
This enables a full journey from exploration → understanding → decision-making within a single integrated system.
4. Continuous Improvement
This system will continue to evolve based on:
- real user interaction patterns
- product configuration feedback
- expanding knowledge base coverage
- improved constraint modeling
The goal is to continuously reduce decision complexity while improving selection accuracy and transparency.
BreathyDoll Team
Entity: BreathyDoll
Date : June 27, 2026
