*This article is approximately 4000 words long and takes about 5 minutes to read.

Introduction: The design center of NetEase Yanxuan is dedicated to providing better products and services for users, allowing them to enjoy the dividends brought by technology and consequently appreciate a better life.

Project Background

AIGC (AI Generated Content) is a new form of content creation. AIGC is becoming increasingly involved in the creative processes of digital content, requiring creators to continuously learn, explore, and innovate to meet the ever-changing market demands.

Exploration of AI Applications in NetEase Yanxuan’s Actual Design Business

NetEase Yanxuan is a lifestyle brand under NetEase, featuring a full range of products, an exclusive app, and a complex and diverse design business environment across all channels. With the continuous development of technology, many changes and opportunities have emerged in the design field. The NetEase Yanxuan design center actively responded by gathering designers to form an AIGC group in early March. This article will detail the applications and explorations of AIGC in various areas such as products, brands, marketing, and experiences from the perspectives of feasibility and promotion, as outlined below:

1) Design ResearchResearch

In the design process, preliminary research is crucial as it helps designers understand client and market needs and trends, guiding design direction and strategy. Designers need to input a large amount of analytical data during the early research phase; for instance, product design requires trend analysis, while experience design needs target analysis, user demographic analysis, and survey analysis. Tools like ChatGPT can quickly gather a wealth of textual information, assisting designers in quickly identifying directions, and also helping them learn methods to solve problems. (Note: This article focuses on the direction of design content production and does not elaborate on ChatGPT‘s research capabilities.)

2) Ideation & ProposalIdeation & Proposal

During the design proposal process, inspiration gathering often generates numerous ideas, but during the conversion of ideas into design drafts, limitations in time and expressiveness may lead to some ideas being discarded. Now, designers can use AI to quickly simulate conceptual creative proposals for presentations and iterate based on demand objectives, responding rapidly to business needs.

In the co-creation project for toilet cleaning, AI was introduced to assist the design team in exploring comprehensive outputs. The design team initially posed questions to ChatGPT from three dimensions: insights into needs, product definition, and how to be credible, and based on the product positioning and channel demands from the BU side, along with the responses from the AI tool, they integrated to derive three clear and defined design directions, ultimately producing packaging designs based on the design strategy. (Note: Marketing concept design strategy – completed by AI, packaging design completed manually.)

3) Photography & Stock ImagesPhotography & Stock Images

AI can assist in shooting and producing high-definition materials, including photography, backgrounds, and image materials. While enhancing the diversity of materials needed for content photography, it also reduces the procurement costs of image materials. For example:

Brand Tone Images

Brand communication requires images that match the brand’s characteristics and style. It is often challenging to find images that fully align with the brand’s style on image websites, causing difficulties in brand communication and necessitating more time and money for on-site shooting. Now, we attempt to use Midjourney (abbreviated as MJ) to generate images that align with the brand’s tone, saving considerable time and procurement costs.

Product Atmosphere Images

In actual shooting, it is necessary to build empty scene settings to capture the required images. Now using MJ tools, we can quickly generate images of empty scene settings with the same theme, and their detail level meets usage requirements, saving time and costs in building scenes and improving production efficiency.

Material Close-ups

In the fashion category, high-definition images showcasing material details are often required, which typically necessitate 3D modeling and rendering (averaging 800-1200 RMB per image). Now, using MJ tools can generate some material images, saving production time and costs while improving efficiency.

Commercial Copyright Images

As an e-commerce platform, a large number of commercial copyright images need to be procured annually for marketing materials (averaging 1500 RMB per image). Now, using MJ tools can generate an unlimited variety of materials as needed, significantly reducing image procurement costs and improving production efficiency.

Clothing Model Outfit Changes

The clothing category requires a large amount of model shooting materials, and AIGC can also solve the issue of character outfit changes. Currently, the restoration degree of products has not reached 100%, but it has shown relevant feasibility. The “virtual characters + clothing change” approach can significantly reduce model and shooting costs for enterprises.

4) Package DesignPackage Design

AI assists designers in providing creative effects in container shapes, graphic design, and backgrounds. As shown below, fast-moving consumer goods packaging designers, after adding keywords, generated different container shapes, graphics, and related atmosphere design directions.

5) Industrial DesignIndustrial Design

AI helps designers quickly verify creative effects. Taking office furniture as an example, models that previously required days to achieve using 3D software can now perhaps be completed in just one day. Although it cannot yet present a finalized solution, it has provided effective materials for initial communication of design direction.

Local new consumer goods designers hope to combine the shape language of tulips with ergonomic chairs to achieve a unique visual effect. Using MJ tools for assistance, they gained a wealth of design inspiration and directions. Additionally, they can test different thematic styles to determine which style best conveys the desired product atmosphere to users.

6) Graphic DesignGraphic Design

AI helps designers generate patterns and apply them to products. Taking kitchen products as an example, it can quickly produce pattern creative directions and scene images.

7) Branding & IP DesignBranding & IP Design

AI helps designers improve the efficiency of searching for materials and modeling rendering. For example, in IP design, designers can use the style tone and content from the preliminary phase as corresponding prompts to produce a large amount of material in MJ. The creative proposals generated by MJ greatly shorten the modeling and rendering process, quickly focusing on the ideal direction.

Through secondary creation by designers, these can be directly applied to proposals for posters, packaging, products, etc., enriching the needs of proposal content. The process from design creative exploration to solution implementation reduces labor costs and enhances the timeliness of design proposals. An application case of “NetEase Tiancheng” pet IP.

Application case of the “Spring Breeze” character IP:

Assistance of AI to the Yanxuan Design Team

1) Assisting Creative Proposals, Enhancing Proposal Efficiency

In open explorations, the group used AI to assist in producing three parts of content: preliminary research, creative design, and high-definition photographic materials. Using AIGC assistance can enhance design quality and efficiency. To help more designers conveniently use AI in design, we have summarized a general design process, which is estimated to help designers of different functions complete complex design proposals in a short time. Meanwhile, the group has also accumulated a large amount of Prompt content as design assets during the exploration process, preparing for future promotional use.

2) Assisting in Photography, Providing Rich Materials

We found that we can generate photography scene materials, props, clothing, characters, etc., around the tone of Yanxuan’s products. The diversity of materials can significantly reduce production costs while improving efficiency and quality. This discovery can help Yanxuan solve some business problems and enhance shooting and production efficiency.

Short-term Focus:AI Assisting in Photography and Reducing Costs and Improving Efficiency

Yanxuan spends a significant amount of time and costs annually on product photography, leading to many problems and difficulties in business operations. Through preliminary explorations, we found that AIGC technology can be used for virtual photography and production, significantly reducing production costs while improving efficiency and quality. Furthermore, the generated high-quality virtual scenes and characters also align very well with the style and requirements of Yanxuan products. To verify the feasibility of this idea, we chose Tiancheng cat food products as experimental subjects, using generated photographic materials to replicate shooting content.

1) Attempting to Use AI to Generate Photography Materials, Enhancing Shooting Efficiency

For example: Designing a material effect image for Tiancheng cat food (see below for the before), requires five steps: scene building & outdoor shooting, model shooting, product shooting & modeling, post-editing, and design production. During outdoor shooting and indoor scene building, there are factors such as long shooting cycles, high shooting costs, and uncontrollable pet models, which may lead to extended shooting times and subpar output quality.

Based on the conclusions from the AI exploration, we attempted to use MJ to assist in the “scene building & outdoor shooting” and “shooting pet model images” stages (see below for the after). Surprisingly, the restoration effect is comparable to photographic materials, and the material scenes are richer and more diverse, with no limit on quantity. After designer layout, the overall quality can meet online requirements. Therefore, we believe the idea of using AI-generated images to replace shooting materials is feasible.

2) Reusing in Broader Product Categories, Observing Generated Effects

Next, we chose the home cleaning category for grounded verification, generating photography images that align with the expected product tone and theme using MJ.

Compared to photographic images, the quality of materials synthesized using MJ has significantly improved.

The effect of product detail pages made using materials generated by MJ also meets online standards.

3) Streamlining Content Production Processes, Establishing Methodologies

From the verification results, using MJ to generate photographic materials for product images is feasible, and the effect meets online standards. Therefore, we believe that the production process for product content can be simplified. Previously requiring five to six steps: “scene building & outdoor shooting, model shooting, product shooting & modeling, post-editing, design production,” can now be reduced to three steps: “still life materials (shooting & modeling), MJ scene synthesis, design production.” This greatly enhances the production efficiency and creativity of product photography content, shortening production cycles.

A New Design Model Beyond Time Constraints Has Emerged

Using AIGC tools to assist in design proposals can improve material collection efficiency, creative efficiency, design quality, and decision-making efficiency, which will have a positive impact on business development, while injecting new vitality into the field of design creativity.

AI assists in photography and production of high-definition materials: “It can provide the necessary material diversity for product photography, save time constraints for shooting, reduce photography costs, lower image material procurement costs, and improve production efficiency.”

Looking to the Future

The rapid development of AI technology also brings opportunities and challenges for digital transformation. We can foresee disruptive changes in virtual streaming content production (such as virtual models, virtual live broadcasts, short videos), voice conversational interactive shopping guides, etc.

The design center is dedicated to providing better products and services for users, focusing on user needs and experiences, continuously researching and exploring how technology can bring convenience and value to users, allowing them to enjoy the dividends brought by technology, and thus more attentively enjoying a beautiful life.

Source of content:Copyright documentation from MJ

MJ Copyright Statement:

Copyright and Trademark

In this section, Paid Member shall refer to a Customer who has subscribed to a paying plan.

Rights You Give to Midjourney

「 Long press the QR code to follow IDC」

「 Long press the QR code to enter the NetEase Yanxuan main site」