Haewon Song

Imagenie

This is a conceptual project designing AI-powered mobile app that enables users to easily create and customize images while maintaining a consistent style. The project involved design research, identifying user pain points, and providing design solutions through wireframing and iterative design refinement.

My Role

UX Researcher, UX Designer

Team

Myself (Self case study)

Year

2024

Duration

3 weeks (Research and Initial Designs)

Problem Statement

AI image generation tools are are experiencing rapid growth with over 34 million images generated daily. However, these tools face critical challenges that limit their usability and potential.

Despite the growing popularity of mobile usage—over 60% of global web traffic now comes from mobile devices (Statista, 2023)—most AI image generators are designed primarily for desktop use, excluding a large portion of users seeking on-the-go creativity.
Many AI tools lack user-friendliness, particularly for those with limited tech experience. Approximately 55% of users report feeling uneasy or confused when interacting with AI tools due to complex interfaces and poor adaptability to individual preferences.
Even when users input identical prompts and settings, AI image generators often produce unpredictable outputs, making it difficult to maintain a cohesive and consistent style. This unpredictability is especially frustrating for users attempting to create images that align with a particular aesthetic or brand identity.

AI image generation tools are are experiencing rapid growth with over 34 million images generated daily. However, these tools face critical challenges that limit their usability and potential.

Despite the growing popularity of mobile usage—over 60% of global web traffic now comes from mobile devices (Statista, 2023)—most AI image generators are designed primarily for desktop use, excluding a large portion of users seeking on-the-go creativity.
Many AI tools lack user-friendliness, particularly for those with limited tech experience. Approximately 55% of users report feeling uneasy or confused when interacting with AI tools due to complex interfaces and poor adaptability to individual preferences.
Even when users input identical prompts and settings, AI image generators often produce unpredictable outputs, making it difficult to maintain a cohesive and consistent style. This unpredictability is especially frustrating for users attempting to create images that align with a particular aesthetic or brand identity.

Problem Statement

AI image generation tools are are experiencing rapid growth with over 34 million images generated daily. However, these tools face critical challenges that limit their usability and potential.

Despite the growing popularity of mobile usage—over 60% of global web traffic now comes from mobile devices (Statista, 2023)—most AI image generators are designed primarily for desktop use, excluding a large portion of users seeking on-the-go creativity.
Many AI tools lack user-friendliness, particularly for those with limited tech experience. Approximately 55% of users report feeling uneasy or confused when interacting with AI tools due to complex interfaces and poor adaptability to individual preferences.
Even when users input identical prompts and settings, AI image generators often produce unpredictable outputs, making it difficult to maintain a cohesive and consistent style. This unpredictability is especially frustrating for users attempting to create images that align with a particular aesthetic or brand identity.

Outcome

Imagenie offers users to easily create customized images with a user-friendly interface that eliminates barriers to AI image generation. With features like the Step-by-Step Tutorial, Inspiration Integration, Partial Generate, and Project Group, Imagenie caters to both beginners and experienced users. These features make the creative process more intuitive, personalized, and consistent.

In the following section, we will dive deeper into how these features work, supported by visual examples that showcase their functionality and impact.

Outcome

In the following section, we will dive deeper into how these features work, supported by visual examples that showcase their functionality and impact.

Step-by-Step Tutorial with Relatable Examples:

Many users perceive AI image generation tools as complex and requiring technical expertise. Imagenie provides an intuitive setup process and tutorial that guides users step-by-step, making the app accessible to everyone even for first-time users or those with minimal exposure to AI technology.

Inspiration Integration from Photo Albums:

Imagenie allows users to upload reference images directly from their device's photo library. transforming everyday photos and screenshots into creative inspiration.

Partial Generate Feature

This feature enables users to upload their own photos and seamlessly blend them with AI-generated elements. By combining real-world images with creative enhancements, users can create unique compositions that reflect their personal style while taking advantage of AI's power for artistic expression.

Project Group for Style Consistency:

Imagenie lets users save generated images into specific project groups, training the AI to understand and replicate a unique style. This ensures that users can generate consistent, cohesive images in future sessions, streamlining their creative process and maintaining a uniform visual identity across their projects.

Process

Persona, User Journey Map, and User Flow

Persona 1 : Emily Carter, 52 years old

Occupation: Handmade Jewelry Shop Owner
Goals:
Create professional-looking product images to enhance her online shop and Instagram presence.
Showcase her handmade jewelry in a way that looks polished and high-quality.
Pain Points:
Limited capital to hire professional photographers or rent studio space.
Finds it difficult to take appealing photos herself due to lack of photography skills and tools.
Unfamiliar with AI and doesn’t know how to give specific prompts to get the desired results.
Has no experience using professional design tools like Photoshop.
Tech Savviness: Basic; familiar with social media apps but not advanced tools.
Quote: "I want my jewelry to look as beautiful online as it does in real life, but I don’t know where to start."

Persona 2 : Daniel Kim, 28 years old

Occupation: Coffee Shop and Bar Owner
Goals:
Promote his events like book talks and small band performances with eye-catching visuals.
Generate unique, original visuals inspired by images he finds online.
Maintain a consistent and recognizable brand style for his shop's social media presence.
Pain Points:
Lacks time to create custom visuals from scratch for every event.
Doesn’t want to simply copy others’ designs but struggles to customize images to his needs.
Wants to ensure his visuals stand out while aligning with his brand.
Tech Savviness: Moderate; comfortable with design apps and editing tools but not highly skilled to create own design assets.
Quote: "I want promotional images that reflect the vibe of my shop—unique, welcoming, and creative—but I don’t have the time or tools to make them."

Persona 1 : Emily Carter, 52 years old

Occupation: Handmade Jewelry Shop Owner
Goals:
Create professional-looking product images to enhance her online shop and Instagram presence.
Showcase her handmade jewelry in a way that looks polished and high-quality.
Pain Points:
Limited capital to hire professional photographers or rent studio space.
Finds it difficult to take appealing photos herself due to lack of photography skills and tools.
Unfamiliar with AI and doesn’t know how to give specific prompts to get the desired results.
Has no experience using professional design tools like Photoshop.
Tech Savviness: Basic; familiar with social media apps but not advanced tools.
Quote: "I want my jewelry to look as beautiful online as it does in real life, but I don’t know where to start."

Persona 2 : Daniel Kim, 28 years old

Occupation: Coffee Shop and Bar Owner
Goals:
Promote his events like book talks and small band performances with eye-catching visuals.
Generate unique, original visuals inspired by images he finds online.
Maintain a consistent and recognizable brand style for his shop's social media presence.
Pain Points:
Lacks time to create custom visuals from scratch for every event.
Doesn’t want to simply copy others’ designs but struggles to customize images to his needs.
Wants to ensure his visuals stand out while aligning with his brand.
Tech Savviness: Moderate; comfortable with design apps and editing tools but not highly skilled to create own design assets.
Quote: "I want promotional images that reflect the vibe of my shop—unique, welcoming, and creative—but I don’t have the time or tools to make them."

Persona, User Journey Map, and User Flow

Persona 1 : Emily Carter, 52 years old

Occupation: Handmade Jewelry Shop Owner
Goals:
Create professional-looking product images to enhance her online shop and Instagram presence.
Showcase her handmade jewelry in a way that looks polished and high-quality.
Pain Points:
Limited capital to hire professional photographers or rent studio space.
Finds it difficult to take appealing photos herself due to lack of photography skills and tools.
Unfamiliar with AI and doesn’t know how to give specific prompts to get the desired results.
Has no experience using professional design tools like Photoshop.
Tech Savviness: Basic; familiar with social media apps but not advanced tools.
Quote: "I want my jewelry to look as beautiful online as it does in real life, but I don’t know where to start."

Persona 2 : Daniel Kim, 28 years old

Occupation: Coffee Shop and Bar Owner
Goals:
Promote his events like book talks and small band performances with eye-catching visuals.
Generate unique, original visuals inspired by images he finds online.
Maintain a consistent and recognizable brand style for his shop's social media presence.
Pain Points:
Lacks time to create custom visuals from scratch for every event.
Doesn’t want to simply copy others’ designs but struggles to customize images to his needs.
Wants to ensure his visuals stand out while aligning with his brand.
Tech Savviness: Moderate; comfortable with design apps and editing tools but not highly skilled to create own design assets.
Quote: "I want promotional images that reflect the vibe of my shop—unique, welcoming, and creative—but I don’t have the time or tools to make them."

Wireframes and design iterations

Based on the personas and the user flow, I developed wireframes to illustrate workflow.

Wireframes and design iterations

Based on the personas and the user flow, I developed wireframes to illustrate workflow.

Wireframes

Reflection

Throughout this conceptual project, I learned valuable lessons about designing for AI-powered products, which significantly different from traditional digital products in several ways.

Designing for Uncertainty
Unlike other digital products, AI-powered tools introduce a level of unpredictability. AI generates images based on the user’s input, which means the outputs can vary each time. This required a focus on guiding user expectations. I had to ensure that users felt in control, even when the results were unexpected or required adjustments. This was achieved by providing feedback on the AI's outcome and giving users clear options to refine their inputs or regenerate results.

Balancing Automation and User Control
Another challenge was determining how much automation the AI should offer versus the control the user should have. In traditional products, we might design for clear user flows with fixed steps. However, in an AI-powered tool, I needed to find the balance between AI-driven automation and the user’s ability to fine-tune results. This balance was key to making the tool intuitive and empowering rather than frustrating.

Iterative Learning and Personalization
One of the most fascinating aspects of this project was integrating the concept of learning from users. AI models evolve as they are exposed to more data, and in the context of a design tool, this creates an opportunity for personalization. The more users interacted with the app, the more their preferences could shape the AI’s future suggestions. I designed the app to offer personalized recommendations and suggested prompts based on a user’s previous creations. This personalized approach made the experience more engaging, but it also highlighted the need for continuous improvement.

For next step, I would improve the product by conducting more user testing to understand how users interpret the AI-generated images and whether they feel in control of the creative process. Testing could focus on the clarity of feedback provided by the AI, the perceived creativity of the generated images, and whether the personalization feels intuitive. Future improvements might include adaptive feedback loops that not only guide users but also allow them to directly influence how the AI learns from their inputs over time.

Throughout this conceptual project, I learned valuable lessons about designing for AI-powered products, which significantly different from traditional digital products in several ways.

Designing for Uncertainty
Unlike other digital products, AI-powered tools introduce a level of unpredictability. AI generates images based on the user’s input, which means the outputs can vary each time. This required a focus on guiding user expectations. I had to ensure that users felt in control, even when the results were unexpected or required adjustments. This was achieved by providing feedback on the AI's outcome and giving users clear options to refine their inputs or regenerate results.

Balancing Automation and User Control
Another challenge was determining how much automation the AI should offer versus the control the user should have. In traditional products, we might design for clear user flows with fixed steps. However, in an AI-powered tool, I needed to find the balance between AI-driven automation and the user’s ability to fine-tune results. This balance was key to making the tool intuitive and empowering rather than frustrating.

Iterative Learning and Personalization
One of the most fascinating aspects of this project was integrating the concept of learning from users. AI models evolve as they are exposed to more data, and in the context of a design tool, this creates an opportunity for personalization. The more users interacted with the app, the more their preferences could shape the AI’s future suggestions. I designed the app to offer personalized recommendations and suggested prompts based on a user’s previous creations. This personalized approach made the experience more engaging, but it also highlighted the need for continuous improvement.

Reflection

Throughout this conceptual project, I learned valuable lessons about designing for AI-powered products, which significantly different from traditional digital products in several ways.

Designing for Uncertainty
Unlike other digital products, AI-powered tools introduce a level of unpredictability. AI generates images based on the user’s input, which means the outputs can vary each time. This required a focus on guiding user expectations. I had to ensure that users felt in control, even when the results were unexpected or required adjustments. This was achieved by providing feedback on the AI's outcome and giving users clear options to refine their inputs or regenerate results.

Balancing Automation and User Control
Another challenge was determining how much automation the AI should offer versus the control the user should have. In traditional products, we might design for clear user flows with fixed steps. However, in an AI-powered tool, I needed to find the balance between AI-driven automation and the user’s ability to fine-tune results. This balance was key to making the tool intuitive and empowering rather than frustrating.

Iterative Learning and Personalization
One of the most fascinating aspects of this project was integrating the concept of learning from users. AI models evolve as they are exposed to more data, and in the context of a design tool, this creates an opportunity for personalization. The more users interacted with the app, the more their preferences could shape the AI’s future suggestions. I designed the app to offer personalized recommendations and suggested prompts based on a user’s previous creations. This personalized approach made the experience more engaging, but it also highlighted the need for continuous improvement.

Go Back To Top