// Cornell · HCI Studio · Product & UX
ApplAI
An applicant-facing platform that makes AI-driven hiring legible: track every application, see how companies hire, and get guidance when an outcome seems unfair.
Overview
Résumés are now read by algorithms first. Applicants rarely know how they were scored, rejections arrive without reasons, and employers decide for themselves how transparent to be. We asked two questions at once, from both sides of the same process.
For job seekers
How might we improve the job search? Help applicants keep track of every application, understand how they are evaluated, and steer toward employers they can trust.
For the hiring system
How might we improve hiring? Make automated decisions more visible and contestable from the outside, without waiting for employers to volunteer it.
What we did
9 stakeholder interviews2 expert consultations2 rounds of usability testing6 core featuresAI-assisted UI audit
My contribution at a glance
Interviews & usability testingFlows & interaction designLo-fi → hi-fi prototypeUI design
Problem
01
Every tool serves the employer
We mapped existing solutions on two axes: who they serve, and whether they aim at technical optimization or at accountability. Employer-side audits are voluntary, regulation like NYC Local Law 144 has limited enforcement, and applicant-side tools only help résumés slip past the filters. One quadrant stayed empty: accountability, designed for the applicant.
Research
02
Five ways hiring fails applicants
We interviewed job seekers, hiring professionals, and industry experts (n = 9), and consulted an AI-law professor and a labor-relations professor. Affinity mapping across all three groups surfaced five recurring breakdowns.
1
Information asymmetry
applicants know ATS exists, not how it weighs them
2
The black box
recruiters distrust AI scores yet still defer to them
3
The feedback void
rejection is an end, not something to learn from
4
The accountability gap
responsibility is spread across vendors and teams
5
Scale amplifies bias
one flawed model can filter people out everywhere
Strategy
03
From optimizing résumés to rebalancing power
Instead of another tool that helps applicants game the filter, ApplAI asks what accountability would look like from the applicant’s side, and sets three objectives that complement regulation rather than wait for it.
Visibility
See the processShow how applications are evaluated, which systems are involved, and what each stage means.
Participation
Apply on fair termsReduce preventable errors and information gaps with preparation tools and a structured workflow.
Recourse
Act when it failsHelp applicants read outcomes, compare company practices, and reach legal guidance when needed.
Design
04
One platform across the hiring journey
The product follows the three phases an applicant actually goes through, so preparing, applying, and evaluating happen in one place instead of across job boards, email, and spreadsheets.
Document reviewpreparation
→
Autofill widgetpreparation
→
Application boardapplication
→
Company evaluatorevaluation
→
Legal support chatevaluation
Testing
05
Two rounds of usability testing
We tested a paper prototype with three job-seeking students and three studio peers, then a high-fidelity Figma prototype in moderated, think-aloud sessions with six active job seekers, five of them international students.
Lo-fi → hi-fi
Screens were too dense, so we cut visual load and clarified hierarchy. The kanban tracker felt familiar, so it became the core. Manual entry was a burden, so we added an autofill widget. Insights became starting points for action, with plain-language explanations of scores.
Hi-fi findings
Centralized tracking gave users a sense of control. But status tags were mistaken for buttons, detail pages lacked a way back, and bias scores needed context. Legal support felt premature for routine applications, while peer comparison and rejection explanations drew the most interest.
An AI-assisted UI audit
To check usability, accessibility, and consistency at scale, we used ChatGPT as an expert-review aid, prompting it to evaluate the main flows against Nielsen’s heuristics, Material Design 3, WCAG 2.2, and the ApplAI brand system.
Nielsen heuristicsMaterial Design 3WCAG 2.2Brand consistency
Takeaway
Accountability can start at the interface.
Applicants cannot see inside every hiring algorithm, but a product can explain where automation is used, record outcomes, and show how employers behave, turning applicants from passive subjects into informed participants.