Atom AI
A design analysis tool.
My Role
Description
An AI tool to address inefficiencies in sourcing design inspiration, detecting plagiarism, and evaluating originality.
duration
3 weeks
Product designer
Tools
Figma
Design inspiration is everywhere—but accessing it efficiently and ethically is harder than it seems. Whether you're a designer searching for resources, a professor grading projects, or an HR professional evaluating portfolios, the challenges remain the same: fragmented tools, time-consuming processes, and limited ways to assess originality.
The design process often involves endless scrolling through Dribbble, inspecting code snippets, or manually comparing submissions. These inefficiencies aren’t just frustrating—they hinder creativity, delay projects, and leave critical gaps in originality checks. As highlighted in The Impossibility of Originality, the debate about whether originality in design can exist underscores the complexities designers face.
Atom AI is a concept project designed to fix this. By leveraging AI to streamline inspiration sourcing, plagiarism detection, and originality scoring.
Context
While a variety of tools exist to assist designers, academics, and HR professionals, many of them fall short of delivering a streamlined solution. We looked at the most common ways they tackle design-related challenges, and turns out, it’s a mixed bag of frustration and workarounds.
Current Methods for Sourcing and Evaluating Designs
Three primary user groups: designers, academics, and HR professionals. These groups were selected based on their direct involvement in the design process—whether it's creating, evaluating, or assessing originality. Over the course of our research, we conducted in-depth interviews and usability tests with 5+ participants from each group to uncover their unique pain points and workflows.
User Stories
Final Designs
A prototype ready for testing
With the design analysis, originality scoring and comparison features in place, we developed a high-fidelity prototype.
Reverse Image Search
User Struggle
Imagine trying to find a needle in a haystack, but the haystack is Google Images, and the needle? It’s an altered, low-res design file.
Key Pain Points:
1
Accuracy is dependent on image quality; low-resolution images yield poor results.
2
Limited scope to conceptually similar designs, missing nuanced overlaps.
3
Doesn’t search in private repositories or unindexed websites
Search web for Image
Manual Inspection
User Struggle
For designers, this means endless scrolling on Behance, inspecting code in websites, or recreating elements from scratch. For professors? Hours of comparing s ubmissions from current and past students.
Key Pain Points:
1
Searching multiple resources is time-consuming and exhausting.
2
Tools are fragmented—no unified way to extract fonts, colors, or UI layouts efficiently.
3
Academics have no clear framework to compare designs or assess originality.
Code Inspection in websites to extract Font details
How it works
1
Design Pattern Recognition: Automatically identifies common patterns and UI elements within a design.
2
Resource Identification: Provides links to design resources, libraries, and related assets for easy access.
4
Accessibility Analysis: Highlights areas for improvement in accessibility using WCAG standards.
5
Comparative Design: Enables users to upload designs or screenshots for side-by-side comparisons.
6
Originality Score: Assigns a unique score to measure how original or inspired a design is.
Atom AI
Iteration
No Context for the Uploaded Files
Final Design
Input fields for categorization of designs for better organization and retrieval
Use case for when the user does not have a design to compare
Iteration
Less visual clarity
Final Design
More intuitive UI cues
Iteration
Images were not clear to the user
Typed out hyperlinks didn’t do much for the user. Feels cluttered.
Final Solution
One platform for design evaluation
An AI powered tool that helps users analyze, compare and verify designs with ease-offering pattern recognition, originality scoring and seamless resource identification all in one place.
Iteration 1
Iteration 2
These stories became the foundation for Atom AI—a tool designed to address their shared frustrations while tailoring features to meet each group’s specific needs.
Problem
Fragmented workflows and inefficiency
Scattered tools and manual processes make sourcing, evaluating, and verifying design originality inefficient, leading to wasted time, inconsistent outcomes, and frustration across the board.
Proposed Solution
Brainstorming our product
With a dinner table sketch we outlined our product. Our solution leverages AI that seamlessly addresses the challenges faced by our users.
Defining our product features
Fulfilling our users needs
Instead of overloading it with complex tools, we prioritized core features for our MVP like efficient design analysis, originality verification, and seamless comparisons
Allows users to analyze fonts
Displays a design’s core components and their sources.
Lists similar designs or resources identified by the AI for inspiration.
Scrollable Interface to organize insights and data for easy navigation.
Initial sketch of our platform idea
Final Design
More visibility
Informing the user where the image was sourced from.
Added a dark layer to highlight the card.
Offering users more options and handing over control.
Iteration 3
Analyzing a design
A unified design analysis solution
Tools
Quickly look up information like icon sets
Comparing Designs
Similar elements are color coded
Putting it in front of our users.
Impact
Measuring impact during the prototype stage was tricky—without a live product, real-world data is limited. To work around this, I simulated usability tests by comparing manual design evaluation vs. Atom AI, showing its potential to save time and streamline workflows even before full implementation.
We conducted a task-based usability test with participants from our target user groups 2 designers, 1 Professor, and 2 HR professionals. Each participant was asked to:
Task 1
Check the originality of a design
Task 2
Identify the font used in the design
They completed these tasks once manually and once using Atom AI.
Task 3
Find the icons used in the design
Task 4
Search for visually similar designs
Outcomes
7X
Total time saved
Atom AI reduced task completion time from 46.5 minutes to 6.5 minutes, saving 40 minutes and making the process 7x faster overall.
6X
Find icons used
Finding icons manually in Figma or searching libraries took 8-12 minutes, but Atom AI detects and links them in 1-2 minutes.
6X
Search for similar designs
Browsing Dribbble, Behance, and Figma took 15-20 minutes, but users were able to find the closest matches in 3 minutes.
5X
Check originality
Atom AI reduced time from 10-15 minutes (Google reverse image search, manual comparisons) to 2 minutes.
Learnings
Simplicity beats complexity
While it was tempting to add multiple features, focusing on a lean MVP ensured that Atom AI solved real user pain points without overwhelming them.
Prototyping is an iterative process
From early sketches to usability tests, every iteration revealed new gaps that we hadn’t considered. This project reinforced that no design is perfect on the first try—user feedback is everything.
Trust is just as important as efficiency
Users appreciated the speed of Atom AI, but they also wanted more transparency in how originality scores were calculated. A great UX isn’t just about fast results—it’s about building confidence in the system.
Shaan Pradhan
Crafted with love and 154 Scorpions songs
Case Study Snapshot
With Atom AI, no more wasting hours hunting for fonts, icons, and similar designs.
Atom AI helps users streamline design evaluation.
Problem
Users struggle with fragmented tools, manual processes, and subjective evaluations when analyzing design originality, sourcing UI elements, and reviewing portfolios. These inefficiencies lead to wasted time and inconsistent results.
Solution
Atom AI simplifies design evaluation by automating originality detection, identifying fonts and icons, and comparing designs side by side.
Key Features
AI-powered originality detection
Automated font and icon identification
Seamless design comparison tool
7X
faster design evaluation compared to traditional methods.
40 mins
saved per task by eliminating manual searching/comparisons.
86%
Faster efficiency for user analyzing design
Atom AI
Smarter Design, Faster Inspiration