← Back to work

Case study, 2024

Walk of Life AI an AI tool for career documents.

A team project designing an AI-powered tool to help students create CVs and portfolios that actually stand out.

Role

UX Research, UI Design Group of 4

Timeline

Semester 1, 2024

Tools

Figma, Zoom Thematic Analysis

Context

University of Melbourne INFO10003

Live prototype — click through it. Open in Figma ↗

01 — The problem

Existing tools make every CV look the same.

Students rely on Canva and Word for CVs and portfolios, but templates are recycled across thousands of applicants. The goal: build an AI tool that helps students move faster without producing generic, AI-flavoured output.

02 — Research

Five students, one big tension.

We ran semi-structured interviews with five undergraduates across design, architecture, data science, sociology, and commerce, then ran a thematic analysis on the transcripts.

"Their templates are quite similar. If everyone applying for the same job is using Canva, then our style is quite similar."

— Participant 1

Six themes

01

Perception of AI

Worry that AI wording reads as generic to employers.

02

Personalisation

Documents need to reflect individual style.

03

Existing tools

Canva and Word — useful but template-heavy.

04

Challenges

Formatting, professional phrasing, time pressure.

05

Uni support

Workshops cover basics but not layout craft.

06

Wanted features

Auto colour coordination and AI detection.

03 — Process

Paper, then pixels.

01

Brainstorming

Mapped each theme to a feature: ease of use → AI chatbot. Need for originality → AI detector. Customisation → drag-and-drop toolbar.

02

Paper prototyping

Each member sketched screens individually; we merged the strongest into a 12-page paper prototype covering the full flow.

03

Figma prototype

High-fidelity build in orange and blue — energy and trust. Component variants let us animate without duplicating pages.

04 — Features

What we built.

Onboarding

AI Chatbot guide

No blank canvas — a chatbot asks for name, field, and style preference, then generates a tailored template.

Authenticity

Built-in AI detector

Scores how AI-generated sections look, and flags wording for humanisation. Tackles the "employers can tell" concern head-on.

Customisation

Drag-and-drop toolbar

Drag in sections, images, and layout blocks. An inline "AI suggested text" bar acts as a writing assistant — not a ghostwriter.

Portfolio mode

Existing-portfolio analysis

Upload an existing portfolio for AI-flagged suggestions and an "Animate with AI" feature that turns static work into interactive visuals.

05 — Testing

Five users, one think-aloud session.

Live testing via Zoom with five participants. 22 tasks each, followed by a short post-trial interview.

5 Participants
22 Tasks per session
~7min Avg. completion

What we learned

"It was very straightforward and to the point, unlike those other websites that do so much unnecessary stuff. This was so easy to use."

— Participant 1

06 — Reflection

What I took away.

The biggest tension was simplicity vs. depth. Efficiency users loved the minimal interface; design-led users wanted more control. In a production version this calls for tiered modes rather than one fixed UI.

Every usability issue we caught — unclear drag-and-drop, overlapping help button, fixed placeholders — surfaced in testing. A reminder of how much value even a small think-aloud round delivers.

The biggest surprise: the ethical dimension. AI can quietly help users misrepresent themselves. Designing guardrails like the AI detector became part of the responsibility, not just a feature.

Next case study

Website Redesign for a Bakery