Designing the Future of Job Search with AI-Powered Recommendations

A UX Research case study on mobile job portals — by Steven Yoshico

Role: UX Researcher Project: AI-driven Job Portal Mobile App (UI/UX Bootcamp — Dibimbing.id, Batch 34) Methods: Quantitative Survey (n=11) · In-depth Interview (n=3) · Affinity Diagram · User Persona · Customer Journey Map Duration: one week

Introduction

Job seeking in the digital age should be easier than ever. Yet for millions of active job seekers — fresh graduates, freelancers, and career switchers alike — the experience remains frustrating: irrelevant listings, confusing filters, and companies that ghost applicants after they’ve put in the effort to apply.

This case study documents a UX research process I conducted as part of a bootcamp mini project at Dibimbing.id. The goal was to uncover what job seekers truly need from a mobile job portal, and how AI can be deployed in a way that genuinely helps — not just as a marketing buzzword.

The findings are structured using the STAR framework: Situation, Task, Action, and Result.

Job seekers are drowning in irrelevant listings — and AI hasn’t helped yet

The global job market is increasingly complex. With hundreds of listings flooding platforms daily, job seekers face a paradox of choice: too many options, too little relevance. AI promised to solve this — yet most job portals still rely on basic keyword matching that fails to account for a candidate’s actual skills, interests, or career trajectory.

As mobile-first job seekers — particularly Gen Z and millennials — shift away from traditional channels (newspapers, career fairs), they lean almost entirely on apps like JobStreet, LinkedIn, and Glints. But these platforms haven’t fully leveraged AI to personalize the experience.

The Numbers Tell the Story

• 73% of users say skill-match with the job listing is the most important factor when choosing to apply

• 7 out of 11 respondents consider job portals highly important in their job search

• 6 out of 11 have stopped using a portal because companies failed to respond after they applied

• More than 50% feel confused and overwhelmed by the sheer volume of listings shown to them

• 7 out of 11 are worried about their data privacy when using AI features

“Companies never responded despite meeting all their ridiculous requirements and qualifications.” — Survey respondent, age 25–30

This is the environment job seekers are navigating every day. The problem isn’t a lack of platforms — it’s a lack of meaningful, personalized, trustworthy experience.

Research goals: Understand what job seekers truly need — and where AI can authentically help

As the UX Researcher on this project, my task was to provide the design team with evidence-based direction for building a new AI-driven job portal mobile app. The research was designed around five core questions:

1. Behavior — How do users currently search for jobs on mobile?
2. Pain points — What are the biggest friction points in existing portals?
3. Feature needs — Which features matter most for finding a job faster?
4. AI readiness — How open are users to AI-driven recommendations?
5. Visual preferences — What kind of interface do users want?

Target Participants

Active job seekers aged 17–30 years, who:

• Use more than 2 job portal apps
• Are smartphone users
• Have basic digital literacy

Mixed-methods research: Survey + In-depth Interviews + Synthesis

To capture both scale and depth, I used a hybrid (mixed-methods) approach.

Quantitative method: Online survey with 11 respondent Qualitative method: In-depth interviews with 3 respondents (~20 minutes each) Synthesis tools: Affinity Diagram, User Personas (×3), Customer Journey Map

Finding 1: User Behavior

Affinity Diagram about user behavior, pain points, features needs

Users are overwhelmingly mobile-first. Smartphones are their primary device for job searching, and they’ve largely moved away from traditional methods like newspapers or physical job boards.

When opening a job portal, their first instinct is to look at salary, location, and job type — not the company name. The most-used platforms are JobStreet and LinkedIn; Kalibrr is the least-used. Most users spend 10–30 minutes per session searching for jobs.

Beyond browsing and applying, users also want to:

• Read company information in depth
• Browse AI-generated job recommendations
• Edit their CV and profile within the app (currently underserved)

Design implication: Salary, location, and job type should be front and center. Filters for these three dimensions must be prominent and highly functional.

Finding 2: Pain Points

Pain points cluster into four categories:

1. Ghosting culture The most universally cited frustration: companies don’t respond after applications. Six out of eleven respondents said this is the primary reason they stop using a platform. The absence of feedback also leaves applicants unable to improve their materials.

“Tidak adanya jawaban sehabis melamar.” (“No answer after applying.”) — Survey respondent

2. Poor filter and search experience Filters are confusing, and more than half of respondents feel overwhelmed by the volume of listings shown without meaningful ranking or personalization. LinkedIn’s filter, in particular, was called out as hard to use.

3. Slow, heavy apps Multiple users complained about slow loading times. The video introduction feature was specifically flagged as unnecessarily cumbersome and performance-draining — users want lightweight UX.

“Loading yang sangat lemot, fitur perkenalan dengan video yang ribet.” (“The loading is extremely slow, and the video introduction feature is complicated.”) — In-depth interview participant

4. Opaque company profiles and scams Vague company descriptions and fraudulent listings seriously erode user trust. Transparency around company information is critical for users to feel confident enough to apply.

“Kurangnya transparansi perusahaan — deskripsi perusahaan sangat singkat, padahal itu sangat penting bagi saya.” (“Lack of company transparency — company descriptions are very short, yet that’s very important to me.”) — Survey respondent

Finding 3: Feature Needs

When asked which features they most need, users ranked them clearly:

• Application status tracking — the most in-demand feature, directly tied to the ghosting problem
• AI job auto-matching — high interest; users want the app to recommend jobs based on their profile automatically
• Job reminders — personalized notifications for new listings that match their background
• CV and cover letter review — suggestions tailored to the specific job being applied for
• Direct recruiter communication — a way to reach a human contact on every listing
• Mini bootcamp / skill development — a surprise insight: users want to stay productive while waiting for responses

“Fitur tambahan yang ingin saya lihat: reminder lamaran yang lebih personal dan disesuaikan dengan pengalaman, skill, serta minat karier pengguna.” — Survey respondent

Affinity Diagram about AI readiness and visual preferences

Finding 4: AI Readiness

This was one of the most nuanced findings. Users are enthusiastic about AI in principle — 8 out of 11 are interested in using AI for job searching — but their enthusiasm comes with conditions.

What users want AI to do:

• Read their profile and automatically recommend matching jobs
• Generate personalized CV and cover letters
• Assist with interview preparation and practice
• Act as a 24/7 career discussion partner

What users are worried about:

• Data privacy (7/11 expressed concern about their data being exposed)
• AI recommendations not being accurate or truly personalized
• Becoming over-reliant on AI outputs that may not reflect their real strengths

“AI helps to define what’s in my head — it was like, already neat and organized. Discussions go faster and it reduces mental load.” — In-depth interview participant, freelancer

Design implication: AI features should be introduced transparently, with clear explanations of how user data is used. The first onboarding experience should build trust before asking for extensive personal information.

Finding 5: Visual Preferences

Users have a clear aesthetic preference: minimalist, clean, and fast.

• Minimalist design is the top-rated visual style
• Search-focused navigation and floating navigation bars are preferred over complex multi-tab interfaces
• Users want a balance of text and visuals — not text-heavy, not image-heavy
• The most important information on a job listing page: job description, salary, and location
• Ease of reading is considered critical by nearly all respondents

“Sederhana, rapi, icon minimalis, warna elegan, dan fiturnya mudah dipahami.” (“Simple, neat, minimalist icons, elegant colors, and features that are easy to understand.”) — Survey respondent

User Personas

Persona 1

Joseph, 27, Freelancer (Yogyakarta)

Joseph is an active freelancer who constantly searches for new opportunities on job portals. He wants AI to match his skills with relevant listings automatically and a direct communication button to reach recruiters. He’s optimistic about AI but frustrated by the silence after applying.

Pain points: No response from companies · Unclear job information · Doesn’t yet know how to use AI features effectively

Needs: AI skill-matching with clear onboarding instructions · Recruiter connect button · Application reminder feature

Persona 2

Mia, 26, Job Seeker (Yogyakarta)

Mia is a social media manager candidate who relies entirely on her smartphone for job searching — she no longer uses newspapers or physical flyers. She is enthusiastic about AI and wants an app that reads her profile and recommends fitting roles automatically.

Pain points: Extremely slow app loading · Video introduction feature too cumbersome and confusing

Needs: Lightweight app performance · AI-optimized profile recommendations · Streamlined application flow

Persona 3

Bagus, 24, Frontend Developer (Yogyakarta)

Bagus is tech-savvy and currently employed, but actively browsing for better opportunities. He navigates straight to salary, location, and job type. He’s frustrated by LinkedIn’s poor filter and by companies that provide minimal information. He envisions AI as a multi-purpose utility button.

Pain points: Sub-optimal AI features · Poor filter experience · Company opacity makes him hesitant to apply

Needs: Versatile AI helper button · Strong filter for salary and location · Mandatory recruiter feedback system

Customer Journey Map: Searching → Applying → Waiting

Phase: Searching → Applying → Waiting Emotion: 😊 Excited → 😤 Frustrated → 😰 Anxious Pain point: App is slow; video intro is tedious; CV and cover letter still fully manual; No feedback; no status update Opportunity: Lightweight UI + AI video intro generator, AI resume & cover letter auto-generation, Mandatory recruiter feedback + live tracker.

Customer Journey Map

A clear product vision: One AI companion that stays with the user at every stage

The research converged on a compelling recommendation: rather than scattering AI features across screens, the app should deliver a persistent One-Stop AI Solution — an always-present AI helper accessible from every screen, capable of supporting the user from first profile setup all the way to post-application follow-up.

Recommendation 1: AI-powered profile matching

Build an algorithm that reads the user’s skills, experience, and career interests to automatically surface the most relevant listings — not just keyword matches. This directly addresses the biggest user goal: finding a job that actually fits.

Recommendation 2: AI resume and cover letter generator

Auto-generate tailored application documents for each specific job listing. This removes one of the most time-consuming friction points in the application process and was among the most requested features across both survey and interview participants.

Recommendation 3: Application status tracker (anti-ghosting)

Implement real-time tracking of every application with mandatory recruiter response requirements enforced by the platform. Users should never be left wondering whether their application was even seen.

Recommendation 4: Lightweight and minimalist UI

Design a clean, search-focused interface with a floating navigation bar. Prioritize fast load times. Remove or radically simplify the video introduction feature. Balance text and visual content. Display salary and location prominently on every listing card.

Recommendation 5: Direct recruiter connect button

Every job listing should include a clear communication touchpoint to a human recruiter. This reduces the most emotionally draining part of job searching — the black-box experience of sending an application and hearing nothing back.

What This Research Unlocked

These findings gave the design team a clear, evidence-based north star: an AI feature set that doesn’t overwhelm users, but instead removes the invisible friction points — the waiting, the manual preparation, the uncertainty — that currently make job searching feel demoralizing.

Users don’t just want a faster job search. They want to feel like the platform is working for them — not just cataloguing listings. AI is the bridge, but only when it’s trustworthy, fast, and genuinely personalized.

“Pengalaman terbaik saya adalah ketika saya berhasil menemukan posisi yang sesuai dengan minat dan kemampuan saya secara lebih cepat dan terarah.” (“My best experience was finding a role that matched my interests and abilities faster and more accurately.”) — Survey respondent

Reflection

This project reinforced something I believe deeply as a UX researcher: data without empathy is just numbers. The survey told us what users wanted. The interviews told us why they wanted it — and more importantly, what the emotional cost of not having it felt like.

The combination of quantitative validation and qualitative depth is what gave the design team not just feature priorities, but a genuine understanding of the people they were designing for.

This case study was conducted as part of the UI/UX & Graphic Design Bootcamp Batch 34 at Dibimbing.id. Research methods included a quantitative survey (n=11) and in-depth interviews (n=3) with active job seekers aged 18–30 in Indonesia. Findings were synthesized using affinity mapping, three user personas, and a customer journey map.

Interested in the full research deck, affinity diagram, or design recommendations? Feel free to reach out.