AI Driven Design -

March . 2025

Taihe Water-Robot

PROCESS HIGHLIGHTS

Project Introduction

Challenge                                                                                    Chance

The challenge is fast iteration under tight deadlines while balancing structure, function, and interaction. The robot must be practical and align with the brand and market, requiring a balance between innovation and feasibility.

This project highlights AI's role in empowering design, offering new insights for future smart service products. AI tools enhance efficiency by integrating data and adapting to the design process.

Timeline

Discipline                              Responsibility                                    Tools

Feb. 2025 - Mar. 2025

Industrial Design

Research
Sketching
Prototyping
3D Modeling
AI-assisted design

Chat GPT
Midjourney
Stable Diffusion
Comfy UI
Rhino
Keyshot

BACKGROUND

Project Background

Taihe Water Group develops smart water solutions. To enhance service and efficiency, it introduced a humanoid robot.As a service and industrial designer, I used AI tools to accelerate the design process, from concept to implementation, ensuring the robot met company needs.The project successfully delivered an intelligent, functional service robot, advancing the company’s smart innovation.

BACKGROUND

AI Work Flow

Step 1 - Mood board creation

In the research phase, we identified three viable design directions through company discussions, market analysis, and competitor studies, ensuring alignment with Taihe Water Group’s brand and smart service needs.

Unlike traditional moodboards, our AI-assisted workflow focused the main board on product form and aesthetics, while supporting images extracted style keywords. AI tools helped refine shape language to match the design vision while maintaining consistency with the company’s product ecosystem, laying the groundwork for detailed refinement and structural optimization.

Moodboard 1 - Hard technology style




Professionalism - technology - future
Emphasizing protection, security, and monitoring
Multi-faceted chest design for a stronger, powerful form

Moodboard 2 - humanity&technology




Friendly - natural - professional
Feminine form, smooth transitions
Rounded, full shape.4o

Moodboard 3 -Natural form aesthetics




Friendly - natural - professional
Natural forms
Streamlined elements
Soft and "cute" aesthetics.

Step 2.1 - Concept & AI-assisted generation

Building on our initial Moodboard, we used MidJourney for base image generation, with supporting images refining prompts and optimizing parameters. This iterative process helped develop concepts aligned with the design direction.

Throughout AI generation, we focused on proportions, form language, feasibility, and brand alignment. After selection and refinement, we ensured the final robot design was both innovative and practical for real-world application.



IMGAE used as AI base reference                                          PROMPT+Base Image generation                                         PROMPT Reverse keyword

AI-assisted concept development

Screening solutions and AI-assisted local modification

AI preliminary concept plan

Step 2.2 - Sketch & AI-assisted generation

In addition to using MidJourney for initial concept generation, we integrated hand sketching and Stable Diffusion for more refined exploration. Sketching allowed us to quickly outline key forms and structural logic, while Stable Diffusion produced high-quality renderings for precise detail control and design expression.

This process enhanced iteration efficiency while maintaining creative control, ensuring a balance of artistic expression and industrial feasibility. It also provided a solid visual foundation for subsequent 3D modeling and structural optimization.

Sketch AI Rendering

AI Preliminary Concept 2

Step 3 - Concept fusion and AI redrawing optimization

After generating multiple concept proposals, we overlaid them with the original model for comparison, analyzing structural feasibility, manufacturing constraints, and functional requirements. Unrealistic elements were eliminated, and the most viable design features were extracted to form an optimized concept.

After generating multiple concept proposals, we overlaid them with the original model for comparison, analyzing structural feasibility, manufacturing constraints, and functional requirements. Unrealistic elements were eliminated, and the most viable design features were extracted to form an optimized concept.



Internal structure of the robot                                                                       Final screening of concept proposals

Merge

AI Repaint

Step 4 - 3D digital modeling

After finalizing the design with Taihe Water Group, we moved to the digital modeling phase, transforming the optimized concept into a precise 3D model to ensure structural and engineering feasibility.

During modeling, we refined proportions, structural details, and material selection, while integrating Taihe Water’s brand identity through form, color, and branding elements. This phase ensured the robot met both visual and functional requirements, laying a solid foundation for prototype manufacturing and testing.

Designing througha different lens.