Vision AI
New approach to AI image generator
What
Why
Role
Category
When
As an enthusiastic advocate of AI-powered solutions, I embarked on this web app design project with the vision of empowering individuals to seamlessly harness the transformative potential of AI in the realm of image creation.
.webp)
Market research
The claim
The AI Image Recognition Market size is expected to grow from USD 4.39 billion in 2023 to USD 10.91 billion by 2028, at a CAGR of 19.94% during the forecast period (2023-2028).
The problem
As AI and deep learning applications expand globally, from tech giants to small and medium-sized businesses, the need for artificial intelligence tools has increased significantly. However, there is a lack of a high-quality image generation program that is both user-friendly and useful.
Competitive analysis
I've completed a comprehensive competitive audit focused on three primary programs—Midjourney, DALL-E 2, and Stable Diffusion—recognized by users as the most prevalent tools in use.
User interviews
I conducted interviews with individuals engaged in the world of AI-powered image generation across social platforms like Facebook and Discord. These conversations revealed a wealth of insights into the current state and future potential of this rapidly advancing technology.
15 participants
Others are the complete opposite, rendering magical images, but having a lousy interface and community.
Charl
The generators themselves are easy to use. but learning how to prompt craft correctly so you get the sort of image you want takes time and has a pretty steep learning curve.
Lynda
Personas
I've meticulously crafted two user personas—the seasoned expert and the eager novice—to capture the diverse needs and aspirations of AI-powered image generation tool users. Experts seek to augment their creative prowess, while novices crave knowledge and exploration. By understanding these distinct perspectives, we can develop tools that cater to both groups, driving effective and user-centric design.
.webp)
.webp)
Armed with the insights gleaned from my thorough research, I seamlessly transitioned to the next phase, devising comprehensive user flows and crafting low-fidelity wireframes to visualize the project's structure and functionality.
%20(1).webp)
Main flow to generate images
To capture the essence of user interactions, I devised a straightforward flow diagram, visually representing the core tasks users can perform. This simplified representation allowed for a clear understanding of the user journey and the necessary functionalities.
It served as a roadmap for the design process, ensuring that each step aligned with user expectations and requirements.
Low-fidelity wireframes
With the flow diagram complete, I embarked on the creation of low-fidelity wireframes, providing a skeletal framework for the user interface design. These bare-bones representations served as a foundation for further iterations and refinements, ensuring seamless user interactions and a cohesive user experience.
High-fidelity UI design
With the initial flow established, I delved into creating a few key screens, starting by defining the fonts and colors. These foundational elements laid the groundwork for a visually appealing and user-friendly interface, ensuring consistent branding and a harmonious aesthetic.
Color palette
Font
High-fidelity prototype
I have crafted a high-fidelity prototype for the visual AI platform, showcasing its ability to generate images based on user-provided prompts. The prototype features large, prominent call-to-action button for intuitive navigation, ensuring a seamless user experience. The interface is designed with simplicity and clarity in mind, facilitating effortless interaction with the AI's image generation capabilities.
.webp)
Design system
I've crafted design elements with atomic components, including interactive buttons and various user interface elements. Below is a fragment of the design system.
Prototype validation
I conducted a usability testing session with four participants, two of whom were experienced users of similar programs and the other two were novices. The primary focus of the test was to evaluate the program's ease of use, particularly regarding prompt entry, resolution adjustments, and prompt saving functionality.
I conducted tests via Zoom calls, introducing the web app prototype to participants and posing questions aimed at assessing their ability to generate images using prompts and their comprehension of the tool's functionalities.
Study results
While (2 out of 4) participants navigated the tool's functionalities seamlessly, the two novice users grappled with the multitude of options, finding it challenging to locate the essential tools required for generating paintings.
Prototype update concept
While time constraints prevented a second usability study on the updated prototype, I incorporated an onboarding process that employs tooltips to sequentially guide users through the purpose of each button, addressing the initial usability hurdles faced by novice users. This enhanced onboarding experience aims to familiarize users with the tool's functionalities and simplify their initial interaction with the application.
Ideas generator
To simplify prompt generation, I introduced an AI-powered tool that generates prompt ideas based on user-provided themes. Users can save or directly apply generated prompts, and the AI provides prompt writing guidance to enhance user experience.
.webp)
This product was an enriching experience. Interacting with users who demonstrated genuine enthusiasm and provided insightful feedback was truly rewarding. Observe the product's progression with new features was an exhilarating process. The field holds immense potential for growth and innovation. As the product gains traction among users, the market demand becomes increasingly apparent, making it an endeavor worth pursuing.
Some of the images I've created have required a hell f a lot of effort and time. The quality of imagery that satisfies some, doesn't satisfy others.
Redican