Project Overview

As part of an AI-driven platform designed to evaluate the ethicality of claims, risk indicators play a critical role in guiding user decision-making. However, accessibility testing revealed that these indicators were not inclusive for colorblind users, creating friction in interpreting risk levels quickly and accurately. Making this core feature accessible was both a usability priority and a trust enabler for the product.

Final Prototype

Industry type: Voice analytics / AI

Company size: 11-50

Project duration: 1.5 weeks

My role: UX & UI Designer

Enhancing Colorblind Accessibility in AI-Driven Risk Indicators

Challenge

  • Color-coded risk states (e.g., High, Low, Average) were difficult for colorblind users to distinguish.

  • Font sizes within indicators were too small, lowering readability.

  • Limited space in the UI meant large visual changes could disrupt layout consistency.

  • Tight resources and time constraints required pragmatic, low-effort solutions that still delivered real impact.

Research & Insights

An assessment was conducted based on WCAG accessibility guidelines and colorblind simulators to replicate user experiences. Key findings included:

  • Poor background/foreground contrast reduced legibility.

  • Shades of green used for multiple indicators were indistinguishable in monochromatic vision.

  • Text inside indicators lost clarity due to small size and low weight.

Visual benchmarking of accessible UI patterns was performed using examples from digital products and adjacent domains such as financial dashboards and stock market interfaces, where fast and inclusive interpretation of high-stakes data is essential.

Key Research Takeaways

  • Patterns/icons can aid distinction but too many variations risk cognitive overload.

  • Increasing typography hierarchy and color contrast improves legibility for all users.

  • Simplicity, familiarity, and consistency are vital for quick recognition.

Screenshot of a project management software showing a risk filter and a list of employees with their roles, dates, and risk assessments marked with color codes for high, potential, low, and average risks.

The background and foreground colors lacked contrast.

Monochromatic view
Screenshot of a project management dashboard with an overall risk filter applied and a list of employees with their roles and dates of employment.

For some, the background colors were faint.

The text on the indicators were too small.

Design Iterations

Iteration One

  • Explored the use of unique symbols for each risk indicator, drawing inspiration from trend symbols (e.g., arrows for stock movement).

  • While icons tested well for clarity, feedback from stakeholders highlighted cognitive load and visual clutter.

Collage of four digital stock market screens displaying stock prices, indices, and charts with various green and red indicators.

Use of simple icons to show trends

Red hospital emergency door sign with upward arrow and letter H
A person riding a bicycle down a street at sunset, with buildings and trees in the background.

High Risk

Yellow parking permit tag with an exclamation mark inside a circle

Potential Risk

Sorry, I cannot determine what is in the image.
A close-up of a red maple leaf with visible veins and a serrated edge.
Green oval traffic sign with a black right arrow and the letter 'A'

Average Risk

A black fork with a silver handle against a white background.
Green button with a white downward arrow and the letter 'L'

Low Risk

Sorry, I can't see the image.
Blue ad badge with check mark

Admission

Please upload the image so I can describe its contents.
DC logo with a stylized infinity symbol in black and white

Disconnected

Sorry, I can't see or interpret the image.
Circle with 'X' and 'NC' inside, label on gray background

Not Connected

Incorporated into the product

There is no image provided to describe.

Iteration Two (Refined Approach)

  • Eliminated unnecessary icons in favor of strong contrast ratios, accessible typography, and subtle background differentiation.

  • Balanced stakeholder priorities (minimal UI disruption) and accessibility principles, creating a lean yet impactful update.

  • Colors were redefined to meet WCAG contrast standards, and font size/weight were adjusted for clarity.

Blank mobile phone screen with a black frame, no content visible.
Red directional sign with arrow pointing up and HR text
Yellow sign with upward arrow and the letters 'PR'
Green button with right arrow and AR letters
Green rectangular sign with a white downward arrow and the letters 'LR'
Blue directional sign with a left arrow and the letters 'AD'
Gray sign with white minus symbol and letters 'DC'
Gray rectangular label with a crossed-out circle and the letters 'NC'

Refined icons with improved consistency

Final Solution

The redesigned risk indicators now enable colorblind users to quickly interpret results through improved contrast, enhanced typography, and clear visual distinctions, without bloating the UI or compromising the list view’s information density.

Screenshot of a web interface showing project results for an engineering hiring process in Singapore, with various language options and access codes.

Monochromatic view

Screenshot of a project management web interface showing testing results for language access codes, with colored labels indicating different statuses for each language.

Standard view

Delivered an accessible, scalable solution within tight constraints.

  • Enhanced clarity and inclusivity of risk assessment, reducing potential user errors.

  • Improved UX for all users, not just those with colorblindness, by strengthening visual hierarchy.

Impact

Reflection

  • Found that focusing on strong color contrast and clear typography is more effective for accessibility without cluttering the interface.

  • Highlighted the importance of balancing accessibility enhancements with spatial and layout constraints for scalable, visually coherent design.

  • Emphasized the need for early stakeholder alignment to prevent fragmented user experiences caused by late design changes.

  • Demonstrated that a simple, user-centered approach to accessibility, focused on clarity and consistency, can lead to effective solutions adaptable to other complex AI platforms.

Thank you for reading!