design case study
Designing AI Models Builder
for Zinier
ROLE
Sr. Product Designer
responsibility
Design research, UX & UI, testing
PROJECT TYPE
UX & UI Designer
ownership
zinier.com
more info
About Zinier
Zinier is non-code customizable field service management platform to schedule and manage workforce and assets.
Abstract
In this case study, I outline the journey of designing an AI models builder that aimed to simplify the process of creating and deploying machine learning models for both technical and non-technical users.
The study covers the challenges we faced, the research and design process we followed, and the outcomes that resulted from an iterative, user-centered approach.
Project Background
A few years ago, I worked on a project to develop an AI models builder—a tool designed to make AI development accessible without requiring deep technical expertise.
This project was born out of the as starting point to make Zinier platform smarter and more automation-driven with AI.
Target Users
- Novice users: Individuals with minimal coding experience who need setup AI models with intuitive interface.
- Developers: Professionals who required a flexible environment model experimentation.
Problem Statement
Technical Complexity
Existing process of building AI Model required extensive coding knowledge.
Resource Constraints and Expensive
Many business lacked the resources to hire specialized AI engineers or training their engineer.
Inefficient Workflows
The process of building AI Model was often fragmented, making experimentation and iteration time-consuming.
Design Goal
The goal was to make the AI Model Builder more accessible for businesses and improving efficiency in training and implementation.
Preliminary Study
AI Model Crash Course
The project started with a deep dive into AI models, as I was unfamiliar how it works and how the data training process. With guidance from Zinier’s AI experts, I spent three days studying the fundamentals while also start sketching potential solutions and wireframe.
What is AI Model?
An AI model is a program or algorithm that has been trained on a dataset to recognize patterns, make decisions, or perform specific tasks without explicit programming. It is the core component of artificial intelligence applications.
In our case, the AI model is built using data collected from the platform, such as task completion times, the number of work orders picked up by technicians, and other operational metrics.
AI Model Crash Course
Research & Discovery
User interview
I began by engaging with a diverse set of potential users—ranging from business analysts to developers—to understand their pain points and knowlege of AI Model. Some key insights included:
- A desire for a visual interface to construct AI models.
- The need for clear feedback during model training and testing.
- Concerns about integrating the AI Model into existing workflows.
Competitive Analysis
I analyzed other AI development tools to identify gaps in the market and understand best practices in UI/UX design for technical platforms.
User interview synthesizing
Design Process
Prototyping
I created low-fidelity wireframes to outline the user flow. These were followed by interactive prototypes that were tested with a small group of users. Some key feedbacks from the user testing:
- There was a big need to streamline AI Model development, especially for people who have low experience in AI and machine learning.
- Need to improved error handling and notifications.
- Need to have visual cues to guide users through complex tasks.
Collaboration with Developers
I worked closely with the engineering team ensured that the design was not only user-friendly but also technically feasible. This collaboration was key in balancing the simplicity of the interface with the advanced functionality required by expert users.
Low fidelity prototype using Zinier existing design components
Implementation & Testing
Some Key Challenges We Cover
Balancing Complexity and Simplicity:
Ensuring that the AI Model builder was powerful enough for experts without overwhelming beginners by utilising template and guided process.
We designed a step-by-step guide for building AI models with a clear, structured flow, along with an AI model builder template that allows beginners to get started without coding tweaks or to build models faster.
Detailed explanations for each instruction to help new users navigate the builder.
At the same time, we still provide options for advanced users to tweak data and customize the template as needed.
Detail matters, we ensure the UI anticipates potential problem that might arise and remains as clear as possible
Plug and play:
The model seamlessly connected to various components within the ecosystem with a single click and publish and it will ready to give recommendation.
Screens
Impact and Lessons Learned
Result:
- Positive User Feedback: Early adopters praised the intuitive design and the ability to experiment quickly.
- Broader Adoption: The tool’s ease of use led to its adoption by teams that previously avoided AI projects due to perceived complexity.
- This tool become a strong direction and foundation for Zinier to focus on automation and machine learning.
Conclusion:
Designing the AI models builder was a exciting project that taught me the value of empathy, iteration, and collaboration. The tool not only addressed a significant market need but also paved the way for more inclusive and accessible AI development practices in the company.
This journey has influenced my approach to design, emphasizing that complex technology can—and should—be made accessible to all. As designers, our role is to simplify complexity for users, ensuring that a tools should remains both powerful and approachable.
Zinier website 2025
Other Works
Redesigning Buy Now Pay Later ExperienceProduct Design
Digital Marketing Simulator Powered by AIProduct Design
Designing AI Models BuilderProduct Design
Wealth Management SystemProduct Design