SeunOlutayo

Equally AI · B2B accessibility platform

Designing an AI-Powered Accessibility Platform at Scale

An AI assistant that turns a wall of WCAG violations into decisions a non-expert can make with confidence and cuts the time it takes a new team to get their first fix live.

Equally AI dashboard showing accessibility health, issues, and remediation guidance

A modern view of the Equally AI product ecosystem, combining scanning, issue triage, and detailed remediation guidance.

Role
Creative / Product Designer
Team
PMs, engineers, marketing, customer
Platform
Web app + browser extension
Tools
Figma, FigJam, Stark, Jira, Notion, Loom
Overview

Making a fast-growing platform feel like one product

Equally AI helps organizations build accessible digital experiences by combining AI-powered accessibility analysis with developer-friendly workflows. The platform enables product teams, developers, and accessibility specialists to identify compliance issues, collaborate across teams, and improve digital products with greater confidence.

When I joined the team, the platform had evolved rapidly, but several parts of the product had become inconsistent as new features were introduced. Navigation, workflows, and visual patterns varied across experiences.

As a Product Designer, I contributed to redesigning the core product experience across multiple initiatives, creating a more cohesive platform while supporting new AI capabilities and enterprise workflows.

My role

Working across the full product lifecycle

I collaborated closely with Product Managers, engineers, marketing, and customers from early discovery through implementation.

  • 01

    Leading end-to-end feature design

  • 02

    Conducting customer discovery interviews

  • 03

    Synthesizing user feedback

  • 04

    Designing complex B2B workflows

  • 05

    Building scalable design system components

  • 06

    Improving accessibility across our own product

  • 07

    Supporting implementation with engineering

Context

Where the work sat

Equally AI detects accessibility issues, explains them, and helps teams resolve them. The product had grown fast, and each new capability arrived with its own patterns. I worked across most of that surface.

Map of the six areas of the Equally AI platform
The six areas of the platform I designed.
The challenge

Simplify accessibility without oversimplifying it

Our users were not casual consumers. Many understood accessibility standards, but translating audit results into actionable improvements still required navigating large volumes of technical information.

Who we designed for

  • Front-end developers
  • Product teams
  • QA engineers
  • Accessibility specialists
  • Enterprise organizations responsible for compliance

What users needed

Understand issues quickly and prioritize what mattered most.

Collaborate with teammates without losing technical context.

Move from detection to resolution efficiently.

The challenge was not adding more functionality. It was making existing functionality easier to understand and use.

Understanding the problem

Finding the friction before exploring solutions

We combined direct customer input with product and business context to understand where the experience was breaking down.

Customer interviewsProduct Manager workshopsInternal stakeholder discussionsFeature feedbackContinuous customer conversationsCompetitive analysis

01

Information overload

Users were presented with large amounts of accessibility data but struggled to determine what required immediate attention.

02

Fragmented workflows

Completing an accessibility audit often required moving between multiple screens, making it difficult to maintain context.

03

Inconsistent experiences

As features evolved independently, similar interactions behaved differently across the product. This increased cognitive load and reduced confidence.

04

Growing product complexity

As AI capabilities and collaboration features expanded, maintaining simplicity became increasingly important.

Design strategy · 01

Simplifying accessibility audit workflows

Detection was never the bottleneck. A flat list of hundreds of findings gave nobody a starting point.

Accessibility audit results shown as one long undifferentiated table
Before: every finding carried the same visual weight, with no clear place to begin.
Equally AI audit results grouped by accessibility requirement with clear priority and status
After: findings are grouped by requirement, with clear priority and next steps.
How might we reduce cognitive load while preserving the technical accuracy expert users depend on?
Design strategy · 02

Modernizing the dashboard

The dashboard became a central hub where users could understand the health of their digital properties. We changed the information hierarchy to surface meaningful insights before detailed data, helping users move from monitoring to action more quickly.

Design strategy · 03

Integrating AI into decision making

Introducing AI here was a trust problem before it was an interface problem. Rather than replacing expert judgement, recommendations acted as guided assistance. Users could understand a suggested improvement before choosing whether to apply it.

An AI suggestion card with a labelled badge, suggested code fix and review controls
Every suggestion is labelled, tied to its issue, and reviewable before anything changes.
Design strategy · 04

Strengthening collaboration

Accessibility is rarely owned by one person. The specialist identifies an issue, the developer fixes it, and QA verifies it. We kept the technical context attached to the issue so teams could communicate progress without rebuilding the story in another tool.

An issue moving from detected to assigned, discussed, fixed and verified across the team
An issue travels from the specialist to the developer to QA, with its status visible the whole way.
Design strategy · 05

Building a scalable design system

Similar interactions behaved differently across the product. That could not be solved screen by screen. I contributed reusable components and interaction patterns that helped engineering deliver features more efficiently while maintaining a consistent experience.

Accessibility audit UI kit with severity chips, issue row variants, detail panel states, button variants, colour tokens and type scale
Severity chips, issue rows, panel states, buttons, colour tokens, and type scale, built as variants the whole product could reuse.
Key design decisions

Learning when not to add more interface

Each decision was made to help users focus on solving accessibility problems rather than learning the interface.

  1. 01Group related information
  2. 02Progressively reveal advanced functionality
  3. 03Improve navigation hierarchy
  4. 04Standardize interactions
  5. 05Reduce visual noise
The product in context

Checking in place

The browser extension puts the audit where the work happens, so a developer does not have to translate a report back onto the page.

The browser extension scanning a live page and highlighting issues in place
The extension scans the page a developer is actually on and pins each issue to the element it belongs to.
Impact

A more cohesive enterprise platform

The redesign contributed to a more cohesive platform that supported experienced accessibility professionals and development teams newer to accessibility practices.

Specific product metrics remain confidential. Success was reflected through positive customer feedback, increased adoption of redesigned experiences, strong collaboration across product and engineering, and continued platform growth.

Equally AI was later acquired by AudioEye, reinforcing the value of the platform and its continued evolution.

Behind the scenes

The work before the final screens

Hand-drawn wireframe sketches of audit and dashboard layouts

Early sketches

Audit and dashboard layouts started as quick FigJam sketches, arguing about what a team should see first before any visual design happened.

Three early concept screens, two of them marked as rejected

Rejected concepts

A single combined issues table, an auto-fix-everything action, and a scoring gauge as the dashboard hero. Each one hid the reasoning users needed to trust the platform.

Component sheet showing buttons, tags, table rows and variant grids

Design system work

Shared components for issue rows, severity tags, filters, empty states and AI suggestion cards, with tokens and variants keeping audits, reporting and collaboration consistent.

Prototype flow map linking connected screens with arrows

Prototypes

Interactive flows covering scan, review, suggestion and share, used in walkthroughs with product and engineering to test the sequence before build.

Reflection

Designing systems, not just screens

Designing for accessibility challenged me to think beyond visual design. Every decision had to balance usability, technical accuracy, business goals, and compliance requirements.

Working on a complex B2B platform strengthened my ability to simplify technical workflows without removing the depth expert users relied upon.

Creating consistent patterns across audits, dashboards, reporting, collaboration, and AI experiences helped make the product feel more intuitive as it continued to grow.