scroll to step aboard
Preeti working on a laptop in a cafe
Preeti
Preeti holding a camera in a pine forest
work in progress in between projects

Hey there, I'm Preeti, |

A product designer with 4+ years of experience turning complex problems into products that people love to use.

Currently at Previously at
reminder: trust the process
scroll to explore
proudly

Selected works

Candidate Journey · Onboarding · Apna

Onboarding
Revamp

Reimagined the candidate onboarding journey, smoothing first-time setup so seekers reach their first meaningful action faster and with less drop-off.

Journey Mapping Flow Design Activation UX
↓
Onboarding
Drop-off
Faster
Time to First
Action
↑
Activation
Rate
Onboarding Revamp
AI Hiring · Employer Platform · Apna

AI Hiring
Assistant

Turned an AI screening SKU into an embedded hiring workflow — making AI explainable evidence recruiters can act on, not a black-box verdict, while keeping them in control of every decision.

UX Strategy Usability Research AI Workflow Design
24×7
Candidate
Calling
~34%
Action on
AI Fit
1.7 min
Time to
Act
Select your AI Agent — the AI Calling Agent set-up step in apnaHire
Job Search · UX/UI · Foundit

Job Listing & JD
Page Revamp

Redesigned the job search results page into a clearer, personalized experience with Match Score insights and richer job details.

User Research Competitive Analysis Iterative Design Usability Testing
36M
Job Views
(from 17.8M)
3.63M
Applications
(from 3M)
+2%
SEO Traffic ↑
Job Listing & JD Page Revamp
0→1 Product · Enterprise AI · Apna

OnlyRounds

Designed a source-agnostic, enterprise-grade AI hiring platform that screens and interviews candidates across every sourcing channel — built solo from concept to launch, with humans kept in control of decisions.

0→1 Product AI Workflow Design Enterprise UX
~80%
Connection Rate
(vs ~40% human)
2×
Recruiter
Productivity
0→1
Platform
Built Solo
05
OnlyRounds · Enterprise AI ATS
Coming Soon
What people say about me

Not just my mom thinks I'm good at this.

"Preeti is born curious and has been very crucial in defining and building user experience at Foundit. She has an eye for detail for the work. She is a go-getter and asks the right questions. She is really quick with iteration and absorbing feedback."

Vijayant Kumar
Vijayant Kumar
Product Manager · Foundit

"Preeti began her career at Foundit, successfully transitioning from an intern to a full-time role. Over her two-year journey, I witnessed her remarkable growth, which gave us the confidence to entrust her with multiple key projects. She consistently delivered outstanding results with minimal guidance, handling them single-handedly."

Thilak Bhat
Thilak Bhat
Senior Product Designer · Cooper

"Had the privilege of working with Preeti at the start of her UI/UX design journey. Her rapid growth and dedication were evident from day one. She has evolved into a skilled designer, showcasing a keen eye for user-centric solutions. Her passion and work ethic make her a standout professional in the field."

Mandeep Vadera
Mandeep Vadera
Design Manager · Foundit

"Preeti is one heck of a hardworking youngster in Foundit's Product Design team. I have literally seen her exponential growth curve before my eyes — from intern to full-time, driven by sheer passion, hard work, and curiosity to improve. She constantly comes up with new ideas and never shies away from a healthy debate."

Saurav Mishra
Saurav Mishra
Product Manager · Foundit

"Preeti is one of those people you can trust with a problem and be confident it will be driven to a thoughtful, high-quality outcome. At Apna, she consistently demonstrated strong ownership, handled complex initiatives independently."

Piyush Tyagi
Piyush Tyagi
Design Lead · Skillz (NYSE: SKLZ)

"Worked with Preeti at Apna and honestly she's one of the best designers I've collaborated with. Her design and product thinking is really sharp — she never rushes to solutions. She digs deep into the problem first, asks the right questions, and challenges assumptions before touching Figma."

Sharath S.p
Sharath S.p
Product Lead · Apna

Get to know me?

View the chapter on my interests, stories
and life outside of work.

Preeti by the sea
Preeti with a plate of street food
Preeti doodling on a giant letter
Preeti — portrait sketch

A bit about Myself✦

the person behind the pixels
Namaste! I'm Preeti ✦

I'm a Product Designer currently designing AI products at Apna, owning end-to-end experiences across candidate and recruiter platforms. I love exploring how design can tell stories, spark emotion, and make everyday moments a little more meaningful.

With 4+ years of experience, my work spans UX audits, research, critique, and prototyping — turning complex problems into intuitive, thoughtful experiences.

My journey started with a love for painting and sketching, and that curiosity still drives how I approach design today — with creativity, empathy, and a genuine interest in people.

Away from the screen, you'll usually find me watching movies, painting, sketching, café hopping, or getting lost in a good series — always collecting little bits of inspiration along the way.

Currently based in Bengaluru, India. If you're building something with heart, let's talk!

Preeti Kundu

My story so far is about learning to see people, patterns, and possibilities, and using design to turn that understanding into something that makes a difference. ✦

User Experience Design
Design Systems
AI-First Design
User Research
Prototyping
UX Audits
UI Design
Usability Testing
Interaction Design

What I bring to the table

Digital experiences that feel effortless, considered, and genuinely useful.

3+ years

Where I've Been

May 2025 – Present
Product Designer II
Apna · HR Tech
Led AI Screening, AI Matching 2.0, and OnlyRounds — a zero-to-one AI ATS platform built solo from concept to launch.
Nov 2024 – Apr 2025
Product Designer I
Apna · HR Tech
Owned recruiter UX — AI mismatch corrections, ECC revamp, WhatsApp Job Invites, database search.
Aug 2022 – Oct 2024
UX/UI Designer
Foundit (formerly Monster.com) · Job Platform
Solo designer for Zuno platform, conversion redesigns, and recruiter surfaces driving 10M+ annual visits.
May 2022 – Jul 2022
Visual Design Intern
Foundit · Job Platform
Campaign pages and onboarding optimisation — contributed to an 8% increase in user sign-ups.
life outside work

A few
snapshots.

Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti working from a cafe
Preeti holding a plate of street food
Suitcases before a trip
Hard Rock Cafe, Bengaluru at night
A wall of illustrated prints
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti working from a cafe
Preeti holding a plate of street food
Suitcases before a trip
Hard Rock Cafe, Bengaluru at night
A wall of illustrated prints
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti at the Under 25 festival
Preeti shooting in a pine forest
Preeti doodling on a giant letter
Preeti at Abbey Falls
Preeti under a pepper-vine canopy
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti
Preeti at the Under 25 festival
Preeti shooting in a pine forest
Preeti doodling on a giant letter
Preeti at Abbey Falls
Preeti under a pepper-vine canopy
Case Study · Foundit · 2023

Job Listing &
JD Page Revamp

Redesigned the existing job search results page, where seekers view job listings and job descriptions, for both mobile and desktop platforms.

My Role
UX/UI Designer
Platform
Web, Mobile, App
Company
Foundit
Year
2023
Final Job Search Result Page

About Foundit: Formerly known as Monster, Foundit is a talent management platform that not only facilitates connections between job seekers and companies but also offers services to enhance candidate skills. It also offers comprehensive solutions for businesses, facilitating the discovery and recruitment of top talent.

What wasn't working

Foundit users face several challenges that hinder their job search experience and diminish engagement:

  • Cluttered Layout: The current design is overcrowded with information, making it challenging for users to navigate and find relevant job listings.
  • Lack of Personalization: The job search results are not tailored to individual user preferences, leading to a less engaging and relevant experience.
  • Fragmented User Experience: Users encounter inconsistencies and interruptions throughout the job search process, resulting in a disjointed experience.
  • Lengthy Job Descriptions: Job descriptions are presented as long paragraphs, making it difficult for users to quickly scan and understand the key details.
  • Insufficient Job Information: Current job cards lack comprehensive information, while also including too many elements, making it difficult for users to quickly identify important information. Additionally, there are no smart insights, such as a Match Score, to help users quickly assess job suitability.
  • Ineffective Filter Functionality: The current filters are dysfunctional and provide irrelevant results, hindering users' ability to refine job searches effectively.
What we set out to do

The primary objective of redesigning the job search results page is to create a streamlined, personalized, and user-friendly experience that addresses current issues of cluttered layout, lack of personalization, broken UX, and insufficient job information. By enhancing visual clarity, incorporating smart insights like a Match Score, and ensuring comprehensive and user-friendly job details, we aimed to significantly improve user satisfaction, engagement, and provide more job applications after the revamp.

Drivers:

  • User Feedback: Users expressed the need for more detailed and easily accessible information about jobs and companies directly from the search results.
  • Business Goals: Increase user engagement and application rates through the platform. Enhance brand loyalty by providing a superior user experience.
  • Technical Feasibility: Integrating new data points like company logos, employee count, and estimated salary involves collaboration with data providers and updating backend systems.

Key Metrics:

  • Engagement: Measure application clicks per search and applications per search session as indicators of improved user engagement.
  • Conversion Rate: Track the increase in applications submitted through the revamped job designs.
  • User Satisfaction: Gather user feedback and ratings on the new job card format.
Job Card Redesign

We began by segmenting the job search result page into three parts: job cards, job descriptions (JD), and filters. Our initial focus was on enhancing the job card section.

Current Job Card Analysis — The current job card included: Job Title, Company Name, Job Type, Job Location, Years of Experience, Salary, Skills, Date Posted, Tag (e.g., Great Place to Work), Save Icon, GPTW Banner.

Current Job Card Analysis

Initial Questions and Considerations:

  • Amount of Information: Do we want this much information on the card? Do users actually read so much information?
  • Skills Display: Should the maximum number of skills be minimized to avoid confusion — ideally between 2 to 4 skills?
  • Company Logos: We should use filled logos for quick recognition, better visibility and layout balance.
  • Job Type Visibility: Is it necessary to show the job type upfront? Shouldn't we cover it in the JD?
  • Salary Representation: Should we shorten the salary display by using units like LPA?
  • GPTW Tag Awareness: Are users aware of the banner? Do they value the "Great Place to Work" tag?
What we learned

To understand how we can enhance the job card experience, we conducted comprehensive user research using a mixed-methods approach including Surveys, In-depth Interviews, and Usability Testing.

Key Findings:

  • Experienced Users (6+ years): Prioritize job title, company name, years of experience required, and salary. Value detailed company information and salary expectations.
  • Intermediate Users (3–5 years): Focus on salary, company name, and job title. Often overwhelmed by too much information and prefer simpler job cards.
  • Entry-Level Users (0–2 years): Look primarily at job title, company name, and years of experience required. Tend to bulk apply but appreciate clear and concise job summaries.
Research iteration 1 Research iteration 2
Redesigned Job Card

The revamped job card includes: Job Title, Location, Experience, Estimated Salary, Company Logo, Smart Tags (Early Applicant, Quick Apply, High Match), Job Posting Date, and Save Job Feature.

Redesigned Job Card

After Redesign: The redesigned job card for the same position maintains the essential details but presents them in a more streamlined and visually appealing manner:

  • Simplified Information: The new design prioritizes key information, making it easier to read at a glance.
  • Location and Work Model: Clearly stated, providing candidates with an immediate understanding of the job's flexibility.
  • Visual Hierarchy: Better hierarchy ensuring crucial information like salary range and experience requirements are prominent.
  • Enhanced Applicant Insights: "Early Applicant" and "Strong Match" badges included, offering candidates a sense of their suitability.
  • Improved Call-to-Action: The "Quick Apply" button remains a focal point, complemented by icons indicating early application and strong match.
Job Description Page

After finalizing the job card, we proceeded to redesign the job description (JD) section. Initial analysis revealed job titles lacked prominence, job descriptions were presented in long paragraphs making it challenging for users to scan, and overall UI improvements were needed.

Current JD Redesigned JD

JD Research Key Findings:

  • Challenging Job Descriptions: Job descriptions are presented in long paragraphs, making it difficult for users to scan. Users prefer bullet points for better readability.
  • Lack of Personalization: Users struggle to understand if the job description matches their profile. Implementing a skill match score can indicate a user's fit.
  • Company Identification: Users prefer seeing company logos as they can quickly identify the company posting the job.
  • Missing Perks and Tags: Users search for perks that are currently missing in job descriptions.
Redesigned JD
Filter Redesign

The existing filter functionality was riddled with inconsistencies, creating confusion for users. Many available filters were underutilized by job seekers. Our goal was to redesign the filter functionality to align with industry standards, simplifying the job application process and enhancing user experience.

Current filters Redesigned filters
Final Job Search Result Page
Final Result Page
Desktop Experience Desktop Experience 2
Results
36M
Job Views
(from 17.8M)
3.63M
Job Applications
(from 3M)
24%
SEO Traffic
(from 22%)
↑
User Satisfaction
Significantly Improved
Case Study · Foundit · 2023

Seeker Profile
Page Revamp

Revamp of the existing seeker profile page of Foundit (formerly Monster.com) to improve usability and significantly boost job seeker engagement, encouraging more frequent and seamless profile updates.

My Role
UX/UI Designer
Scope
Usability Testing, Survey, Research
Timeline
1 Week Research
Participants
6 Users
Seeker Profile Cover

About Foundit: Formerly known as Monster, Foundit is a talent management platform that not only facilitates connections between job seekers and companies but also offers services to enhance candidate skills. It also offers comprehensive solutions for businesses, facilitating the discovery and recruitment of top talent.

AIM of the Project

The current seeker profile isn't user-friendly and has a fragmented, broken experience. This issue is adversely affecting the frequency of profile updates.

By revamping the profile page, we aimed to improve usability and significantly boost job seeker engagement, encouraging more frequent and seamless profile updates. Our goal was to create an enhanced and intuitive user experience that increases user interaction and satisfaction.

What success looked like
  • Increase Profile Updates: Drive website traffic to the profile page to boost updates and reduce drop-off caused by poor user experience.
  • Improve Customer Satisfaction: Optimize visual hierarchy, personalization, accessibility, integration, and security to close UX gaps and deliver a cohesive experience.
  • Enhance User Engagement and Minimize Drop-offs: Implement interactive features, encouraging job seekers to spend more time on the platform.
Before Redesign

Current M-site and App experience showing the pain points:

Current App Current M-site
Usability Study

Participants: 6 users, mixed gender, age 22–34 years. We conducted moderated usability testing over 1 week.

Tasks given to participants:

  • Task 1 — Profile Update: "Imagine you want to update your profile information. Can you please walk us through how you would attempt to do that on the current profile page? Share your thoughts out loud as you go through the process."
  • Task 2 — Navigation: "Let's explore the current profile page together. Look for information about your work experience, skills, and education. Tell us how easy or challenging it is to find and update these details."
  • Task 3 — Feedback on Usability: "Share your thoughts on what aspects of the user interface you found user-friendly or challenging? Are there specific features that stand out to you, positively or negatively?"
  • Task 4 — Proposed Changes: "Based on your experience, what improvements would you suggest to make it easier for users to navigate and update their profiles? Are there any features you think should be added or removed?"
What we found
  • Reduce Pop-ups (6 people): Limit the number of pop-ups to minimize user irritation and provide a smoother browsing experience.
  • Emphasise Profile Score (4 people): Highlight the profile score for increased visibility, especially for users returning after a while, to encourage engagement in profile completion.
  • Optimize Resume Upload (4 people): Ensure that users' information is parsed effectively after uploading a resume to streamline the profile completion process.
  • Minimize Banner Prominence (4 people): Decrease the prominence of banners to avoid user confusion with ads, enhancing overall user experience.
  • Improve Mobile Experience (3 people): Address issues with mobile pop-ups and broken mobile number pop-up to enhance the mobile user experience.
  • Prioritize Work Experience (3 people): Recognize the importance of work experience for users; consider adjusting the visual hierarchy to give it more prominence.
  • Enhance Skill Visibility (2 people): Place skills higher in the visual hierarchy and consider asking for skills first to cater to user preferences.
  • Multiple Resumes Option (1 person): Introduce the option for users to maintain multiple resumes, accommodating diverse job applications.
Persona 1 Persona 2
Who we designed for

Priyanka Sharma, 32, Finance Manager — Main Goal: Secure a role with ample growth opportunities. Pain points: Annoyed by excessive pop-ups disrupting smooth profile navigation. Believes the profile score is not easily visible. Finds prominent banners distracting, mistaking them for ads. Finds it difficult to navigate between different sections.

Ansh Thakur, 25, Software Engineer — Main Goal: Obtain a senior engineering position at a leading tech company. Pain points: Frustrated by numerous pop-ups that interfere with seamless profile navigation. Finds the navigation between profile sections cumbersome and time-consuming. Notices inconsistencies in the design that make the user experience less professional.

Current state data

Avg traffic on profile page (Sep 23 – Jan 24):

  • Desktop is used more to visit profile — 2.5M avg; engagement time per session: 1min, 16sec
  • More traffic from emailers — 1M avg; engagement time per session: 1min, 12sec
  • Desktop is more used from emailers; however engagement time is more when user comes from Android — avg engagement time: 2min, 07sec
  • High drop-off rates observed across all platforms
Data insights
Redesign priorities

The user feedback analysis highlights crucial redesign priorities for the job seeker profile page. Streamlining intrusive pop-ups, prioritizing seamless mobile access, and enhancing profile score visibility are focal points. Addressing banner inconsistencies and emphasizing the work experience section align with user preferences. The redesign will integrate robust resume parsing, clear prompts, and early skills placement in the visual hierarchy. Introducing multiple resumes and refining job recommendations aim to accommodate diverse needs. The overarching goal is a user-friendly, visually cohesive profile page tailored to the unique preferences of job seekers in India, fostering a positive and efficient user experience.

The redesign
Redesigned Profile Redesigned Profile 2
Product · Enterprise AI · 0→1

OnlyRounds

A source-agnostic, enterprise-grade AI hiring platform that screens, interviews, and coordinates candidates across every sourcing channel, while keeping humans in control of the final decision. Built solo from concept to launch.

My Role
Product Designer · 0→1
Timeline
2025
Team
Solo Designer
Platform
Enterprise Web · Recruiter Console
🖼️
Hero: OnlyRounds recruiter console
Drop the cover render here (Figma: OnlyRounds dashboard / job lifecycle). Replace this block with an <img>.
About

An AI hiring layer that sits above every sourcing channel.

Enterprise hiring teams handle huge volumes of candidates from many channels at once: job portals, ATS, databases, referrals. OnlyRounds is a source-agnostic, enterprise-grade AI screening and interviewing platform that lets teams screen, interview, and coordinate candidates consistently using AI agents, no matter where those candidates came from.

What it is not: not limited to apna-sourced candidates, not a single static chatbot, and not a hiring-decision engine. Final calls always stay with humans.

Context

Single-channel AI screening couldn't scale to the enterprise.

apna's earlier AI screening product (AISKU) worked well for SMBs because apna was their primary sourcing channel. For enterprises, apna contributed only a fraction of total sourcing, so most candidates still fell back to slow, manual screening and the ROI of automation collapsed.

10-20%
Share of enterprise sourcing that came from apna. The rest was screened manually.
1 channel
Existing AI tools automated only a single channel, leaving partial coverage and poor ROI.
4+ sources
Portals, ATS, databases, referrals, each with its own format and candidate state.
~40%
Human recruiter connection rate. Most outreach never reached the candidate.
The Challenge 🔎

Screen every candidate the same way, no matter where they came from.

Design a platform that unifies all sourcing channels into one pipeline, lets recruiters configure trustworthy AI agents for screening and interviewing, and orchestrates outreach at enterprise scale, all without taking the hiring decision out of human hands.

🧭
Before: fragmented, multi-channel screening
Figma: current-state map of siloed channels.
My Role 🧩

Sole designer, end to end, concept to launch.

01
Area
Job Lifecycle & Setup

JD ingestion, structured job creation, interview-round setup, and evaluation-criteria editing.

↳ Turned a messy JD into a structured source of truth.
02
Area
AI Agent Configuration

Configuring voice/video agents, language, persona, and reviewable scripts, with a test-before-go-live step.

↳ Made complex agent setup reviewable and safe.
03
Area
Candidate Pipeline & Flows

Multiple entry points converging into one pipeline, with distinct inbound and outbound (consent-first) experiences.

↳ One pipeline, two candidate realities handled.
04
Area
Recruiter Dashboard & Reporting

Status tracking, recordings, AI summaries, and credit/usage reporting for the workspace.

↳ Decisions stay with humans, backed by evidence.
Design Goals 🎯

Three questions framed the whole platform.

?
How might we
Unify candidates from every sourcing channel into a single, consistent pipeline?
?
How might we
Let recruiters configure trustworthy AI screening & interview agents without drowning in complexity?
?
How might we
Automate outreach and coordination at scale while keeping humans in control of every decision?
Solution Walkthrough 🎨

Four surfaces, one end-to-end hiring flow.

Surface 01 · Job Setup

From JD to a configured, testable job

Recruiters paste, upload, or AI-generate a JD; an LLM extracts and cleans it into a structured job that becomes the source of truth. They then set up interview rounds (screening, interview, or human-led tasks), edit AI-generated evaluation criteria, configure each agent (voice/video, language, persona), and test the agent before going live.

⚙️
Job setup & round configuration
Figma: JD → structured job → rounds → agent config.
Why this improves UX

A single source-of-truth JD plus editable, AI-generated criteria makes setup fast without giving up recruiter control, and the test step builds trust before a real candidate is ever called.

Surface 02 · Candidate Pipeline

Every source converges into one pipeline

Candidates enter via bulk upload, embedded apply links, direct interview links, or auto-sourcing from ATS and job platforms, and all of them land in a single pipeline per job. Inbound candidates (who have intent) go straight into screening with a short intro; outbound candidates (contacted from a database) get an interest check and give consent first. The screening criteria stay identical across both; only the entry script changes.

📥
Unified candidate pipeline
Bulk upload / links / ATS → one pipeline.
🔀
Inbound vs outbound flows
Consent-first entry for outbound.
Why this improves UX

Keeping evaluation criteria constant while varying only the entry script gives fair, comparable results across sources, and consent-first outbound respects candidates who never asked to be contacted.

Surface 03 · Orchestration

The workflow decides when, how, and what next

An orchestration engine coordinates multi-channel outreach, deciding when to contact a candidate, through which channel (call, WhatsApp, email), for what objective (interest, screening, interview, scheduling), and what to do on response or silence. It optimises for time-of-day and honours retry limits and night-time policies. The design principle: agents decide what task; the workflow decides when and how.

🕸️
Workflow orchestration
Channel · timing · retry logic.
Why this improves UX

Separating "what" (agents) from "when/how" (workflow) kept the system understandable to configure and dramatically lifted reach; automated coordination hit candidates at the right moment on the right channel.

Surface 04 · Evaluation & Dashboard

Evidence in, clear status out

After each interaction, the data is scored by an evaluation agent against the recruiter's criteria and resolved to a clear status: Fit, Not Fit, Partial Fit, Not Interested, No Response, or In-Progress. In the recruiter portal, teams see status at a glance, listen to recordings or watch videos, and read AI-generated summaries, plus credit and usage reporting for the workspace.

📊
Recruiter dashboard & candidate status
Status · recordings · AI summaries.
Why this improves UX

A tight set of well-defined statuses plus one-click access to the recording and summary lets recruiters act in seconds, and keeps the final judgment human, backed by the actual evidence.

Impact 📈

Enterprise AI hiring, orchestrated at scale.

~80%
Connection Rate
(vs ~40% human)
2×
Recruiter
Productivity
All
Sourcing Channels
Unified
0→1
Platform
Built Solo
  • Roughly 2× the connection rate of human recruiters through workflow-driven, multi-channel outreach.
  • 2× recruiter productivity by automating coordination, screening, and first-round interviews.
  • Consistent, unbiased evaluation: the same criteria applied to every candidate, from every source.
  • Extended AI hiring from SMBs to the enterprise by becoming source-agnostic.
Key Learnings ✨

Reflections from building an enterprise AI product 0→1.

1

Separate "what" from "when"

Letting agents own the task and the workflow own timing and channel kept an inherently complex system configurable and legible.

2

Source-agnostic is a design constraint, not just a backend one

Every source has a different candidate state; the UI had to normalise them into one pipeline without hiding what mattered.

3

Trust comes from control + evidence

Editable criteria, test-before-go-live, recordings, and summaries let recruiters hand work to AI while keeping the decision, and the accountability, human.

4

Consent is part of the experience

Outbound candidates never asked to be contacted, so an interest-check-first flow wasn't a nicety; it was the design.

0 to 1 Project

AI Hiring Assistant

Recruiters hiring frontline roles receive hundreds, even thousands, of applications per job. AI Recruiter interviews candidates, evaluates answers against recruiter-defined criteria, summarises every conversation, and recommends qualified candidates before any human interaction. It was never meant to replace recruiters. It removes repetitive first-round screening while recruiters keep complete control of the final decision.

My Role
Product Designer · 3 of 4 Verticals
Timeline
4 Months
Team
PM, Engineering, AI, Business
Platform
Employer Web · Candidate App

The shipped product in action: AI Agent evaluation inside the recruiter dashboard.

AI Calling Agent live in the apna recruiter dashboard
/ 01 Introduction

An AI recruiter that completes the first screening round before you even open the dashboard.

apna is India's largest hiring platform for frontline roles: telecalling, delivery, customer support, sales. Volume is the defining problem. Every candidate still needs manual screening, repetitive qualification checks, and follow-up calls before a recruiter can identify the right fit.

AI Hiring Assistant exists so that hours of daily calling stop delaying hiring. The AI conducts the first screening interview automatically. Recruiters open the dashboard to qualified, explained recommendations instead of a raw pile of applications.

/ 02 My Role

I owned 3 of the 4 verticals, end to end.

AI Screening Setup Employer Candidate Connect Candidate Experience Pricing & Monetization · Monetization team

Pricing and Monetization was owned by the Monetization team. I collaborated closely with them, along with Product Managers, Engineers, AI teams and Business stakeholders, through research, design, validation and launch.

/ 03 Timeline

Four months, research to launch.

Research→Current workflow→Problem definition→Concept exploration→Iteration 1→User testing→Research insights→Iteration 2→Launch→Measure→Phase 2

This chapter covers the journey up to Iteration 2. Launch results and Phase 2 come later.

/ 04 The Problem

Recruiters spend more time filtering than hiring.

Post job→Hundreds apply→Open every profile→Call individually→Ask the same questions→Filter manually→Schedule interviews

Recruiters hiring at scale repeatedly told us the same thing:

"I spend more time filtering candidates than actually interviewing them."

4 to 6 hrs
A recruiter calling 50 applicants a day spends this just to find 3 worth interviewing.
48 hrs
The best applicants get offers within two days. Slow screening loses them.
100×
The same five questions, notice period, availability, CTC, asked over and over.
1 lang
English-first tools mis-ranked candidates who were perfectly fluent in Hindi.
The opportunity

Instead of helping recruiters screen candidates faster, what if the first screening round was already completed before recruiters even opened the dashboard? That became the foundation of AI Recruiter.

/ 05 Goals

One product goal. One business goal. Four things in balance.

Product goal
  • Automate the first round of hiring interviews.
  • Configure interviews in minutes, not hours.
  • Keep the candidate experience conversational.
  • Reduce recruiter effort without removing control.
Business goal
  • A premium offering that grows revenue.
  • Differentiate apnaHire from traditional job boards.
  • Success meant balancing four forces: AI adoption, recruiter trust, candidate completion, monetization.
/ 06 Success Metrics

Product-level outcomes, not screen-level wins.

  • Increase AI Recruiter adoption.
  • Reduce manual recruiter screening effort.
  • Improve recruiter confidence in AI recommendations.
  • Increase premium attachment rate.
  • Improve qualified candidate conversion.
/ 07 Problem Statement

Recruiters did not want AI to replace their judgment.

They wanted it to remove repetitive work while giving them enough evidence and control to make the final decision.

How might we

Help recruiters use AI screening to make faster hiring decisions while preserving trust, control, and confidence?

/ 08 The Current Hiring Workflow

We mapped both journeys before designing anything.

The legacy flow, end to end, from checkout to candidate list:

Legacy recruiter and candidate journey on apnaHire

The recruiter journey

Purchase AI Recruiter→Configure screening→AI interviews candidates→Review shortlisted→Contact qualified→Hire

The candidate journey

Apply→Receive AI interview→Complete interview→AI evaluates→Recruiter reviews→Interview

Mapping this made one thing obvious: AI Recruiter was not a single feature. It touched multiple products across the hiring ecosystem.

/ 09 Product Scope

Four connected touchpoints.

01
Vertical
Pricing & Monetization

Introduce AI Recruiter as an add-on while making its value immediately understandable.

02
Vertical
AI Screening Setup

Help recruiters configure interview criteria with minimal effort.

03
Vertical
Employer Candidate Connect

Present AI recommendations in a way recruiters trust and can confidently act upon.

04
Vertical
Candidate Experience

An interview flow that feels conversational rather than intimidating.

/ 10 Phase 1

We started lightweight on purpose.

Rather than attempting to perfect the experience immediately, we built a set of initial concepts covering all four touchpoints. The goal was not visual polish. The goal was learning, before investing heavily in development.

?
Comprehension
Will recruiters understand AI Recruiter?
?
Trust
Will recruiters trust AI recommendations?
?
Setup
Will recruiters understand screening setup?
?
Candidates
Will candidates understand the AI interview?
?
Value
Will recruiters see enough value to pay for it?

These became our first hypotheses. The early explorations below became our Phase 1 prototype.

/ 11 Phase 1 · Iteration 1

Deliberately lightweight. Built to be tested, not shipped.

Iteration 1 · Vertical 01

Pricing & Monetization

Hypothesis: recruiters will pay for AI at checkout if the value is obvious at the moment of posting a job.

We slotted a Smart-AI tier beside Classic and Premium, priced above Premium, with an inline demo so recruiters could experience a screening call before paying. What we expected to learn: whether checkout is where conviction happens.

Iteration 1 pricing: Classic, Premium and Smart-hire tiers with checkout

The first pricing concept: a third AI tier with a demo interview entry, straight into checkout.

Demo interview modal with Monika AI

The demo: recruiters talk to the AI before they buy it.

Iteration 1 · Vertical 02

Screening Question Setup

Problem: configuration effort kills adoption. Assumption: if AI drafts the questions from the job details, recruiters only need to review, mark deal-breakers, and continue.

AI recommended screening questions with deal-breaker toggles

Questions pre-drafted by AI from the job post. Deal-breaker toggles make criteria decisive.

Editing a screening question with answer types

Full control on tap: answer types, preferred answers, remove.

Job live confirmation leading into screening setup

Setup opens right after the job goes live, while intent is high.

Expected behaviour: recruiters accept most AI questions, edit a few, and finish setup in minutes.

Iteration 1 · Vertical 03

Employer Candidate Connect

The riskiest questions lived here. Cards or tables? Depth or speed? Rather than commit, we explored four dashboard concepts and let the trade-offs surface. No final direction yet. The purpose was divergent thinking.

Rich cards with static screening panel

Rich cards: maximum transparency per candidate.

Table view with score summary

Table: maximum scanning speed.

Master detail split view

Master-detail: list beside full AI evidence.

List with filters sidebar

Filter-first: control for high-volume pipelines.

Concept
Trade-off
Rich cards
Clear outcomes per candidate, but tall and repetitive at volume.
Table view
Fastest triage, but scores feel generic without visible proof.
Master-detail
Evidence beside the list, but crowded on small laptops.
Filter-first
Powerful drill-down, but choice overload before value.
Iteration 1 · Vertical 04

Candidate Experience

Approach: conversational, not intimidating. The AI screening call reached candidates right after applying, with a simple in-call layout: duration, mic and volume controls, nothing to learn.

Candidate mobile flow with AI screening call

Iteration 1: the call arrives immediately after applying, while intent is hot.

We wanted completion and comfort. Whether an instant call delivers either was exactly what testing had to answer.

/ 12 User Research & Validation

Then we watched real recruiters work.

With Iteration 1 in front of recruiters, we needed evidence, not opinions. We wanted to understand how recruiters actually screened, what information they trusted, which UI elements they ignored, whether AI influenced decisions, and where friction lived.

Contextual inquiry Shadowing recruiters Task-based usability Observation AI and non-AI recruiters
01 · Trust
Trusted AI, verified anyway
Recruiters re-checked questions and details, and called every candidate regardless of score. Opportunity: make verification effortless instead of fighting it.
02 · Visibility
Hidden insights are ignored insights
Insights two clicks deep were opened 3 times in 10. Opportunity: surface AI evaluation where the decision happens, one click at most.
03 · Speed
Speed beat information density
Lists were scanned in seconds; details opened only when necessary. Opportunity: rapid scan first, depth on demand.
04 · Friction
Manual actions broke the flow
Typing numbers by hand, multi-step status changes, constant callbacks. Opportunity: one-tap call, WhatsApp and status actions.
05 · Signals
Five signals drove every decision
Location, salary, education, experience, communication. Opportunity: pin these to the surface of every candidate.
06 · Proof
Audio was proof, not primary
Recordings were played only when AI conflicted with expectations. Opportunity: keep them one tap away as the tie-breaker.
/ 13 Design Principles

Six rules, straight from the research.

01
Principle
Surface AI insights where decisions happen
02
Principle
Support rapid scanning before deep analysis
03
Principle
Keep recruiters in control
04
Principle
Make AI explainable, not authoritative
05
Principle
Reduce repetitive manual work
06
Principle
Design for high-volume hiring workflows
/ 14 Phase 1 · Iteration 2

Not a new design. An evolution, with evidence behind every change.

Iteration 2 · Pricing

Value before the paywall

Insight: AI was not convincing as a standalone purchase; recruiters wanted to experience it first. Problem: the demo lived too deep in checkout. Change: the AI card moved to the centre with a Recommended tag, value props on the card, and the AI recruiter introduced on the dashboard itself, before posting, with a walkthrough and real voice clips. Expected impact: higher demo engagement and premium attachment.

Centered Smart AI card with Recommended tag

The AI tier moves centre. Framing does the selling.

Smart AI card with waveform and value props

Value props and the voice waveform, on the card itself.

Dashboard banner introducing the AI recruiter

Meet your AI recruiter, before you even post.

Walkthrough popup and voice selection

The walkthrough, plus selectable pre-recorded voices.

Iteration 2 · Setup

Configuring an employee, not filling a form

Insight: recruiters said setup felt like configuring an employee. They wanted to hear the voice and control the company pitch. Problem: Iteration 1 read as a form. Change: voice avatars in English, Hindi and regional languages, an editable company pitch, and AI-generated questions to review. Expected impact: setup completion up, drop-offs down.

Agentic setup with voice avatars Monica, Diya and Tanvi

Pick a voice, shape the pitch, review the questions.

Compiling and success states

Progressive feedback while the agent gets ready.

Iteration 2 · Dashboard

Evidence at the surface

Insight: hidden insights were ignored; five signals drove every decision. Problem: reasoning sat two clicks deep. Change: an AI Agent Fit tab in the pipeline, with the evaluation, criteria checks and key signals visible directly on the candidate, and recordings one tap away as the tie-breaker. Expected impact: faster shortlists, higher trust in Fit.

AI Agent Fit tab with visible evaluation in the dashboard

The direction after testing: a fast list, AI evaluation at the surface, proof one tap away.

Iteration 2 · Candidate

A breath before the interview

Insight: instant calls made candidates anxious; they wanted preparation time. Problem: the interview started the moment they applied. Change: a transition state, "your interview is about to begin", with preparation cues before the AI connects. Expected impact: higher completion, lower drop-off.

📱
Interview transition screen
Screen pending export: the pre-interview preparation state.
Next

Iteration 2 shipped. What follows is what six months of live numbers told us.

The design was shipped after minor back and forth with engineering and PMs
As usual !! :)
Now comes the fun part
Did users actually use it?
/ 15 Launch & Early Results

Phase 1 went live. Then we measured everything, weekly, for six months.

Every number below comes from the live metrics tracker, from launch week through the end of January.

1 in 10
Explorers posted
an AI job (peak 15.8%)
~100%
Of leads screened
automatically
1 in 5
Candidates
tagged FIT
48%
Recruiter action on
AI Fit, launch week
  • Recruiters who explored the AI SKU converted to posting an AI job at roughly 10% on average, peaking at 15.8% in the early weeks.
  • The agent screened a median of 9 to 10 leads per job per day, with screened volume tracking total volume almost exactly: the first round genuinely ran itself.
  • About 20% of screened candidates were tagged FIT, with 12 to 14% Full FIT: a strong, consistent qualification signal.
  • Inconclusive evaluations fell from 27% to about 10% across the period as screening and prompts improved.
  • Jobs hit 88 to 91% of their target lead fulfilment at expiry.
  • Discovery was the real top-of-funnel constraint: only 5 to 6% of recruiters shown the banner went on to explore the AI SKU.
  • Early repeat purchase appeared among existing users, though cohorts were still small and volatile.

And the numbers that told us the truth:

What worked
  • Launch-week engagement was strong: recruiters acted on 48% of AI Fit leads within 3 days.
  • Screening quality kept improving month over month.
What warned us
  • Action on AI Fit leads decayed to about 29 to 31%, with a median 20 hours to first action.
  • Candidate no-response climbed toward 35 to 44% as unknown-number fatigue set in.
  • Shortlist rate stayed at 4 to 6%: recruiters still verified everything themselves.
/ 16 What Phase 1 Taught Us

Every number and complaint pointed at something fixable.

We read the data and the complaints case by case. Five fixes shipped while Phase 1 was still live.

01
Candidate drop-offs
Interviews had no memory

Mid-interview drop-offs died as "inconclusive". Restarting meant repeating everything.

Agent memoryResume where you leftMulti-call interviewsRecording per call
↳ Inconclusive fell 27% to ~10%. And 15 to 22% of Potential FIT upgraded on a later call.
02
Reachability
Unknown numbers get ignored

No-response climbed toward 35 to 44%.

Smart retriesWhatsApp resume linkIn-app nudgeCall me later
↳ The interview no longer depends on one perfect pickup.
03
Multi-job candidates
Same call, different jobs

Candidates with several AI applications could not tell the calls apart.

Job + company introQueued interviews
↳ Every call now announces which job it is for.
04
Screening strictness
One size fits no one

Some recruiters wanted only top matches, others a broad pool.

High PrecisionBalancedBroad Pool
↳ Strictness is editable mid-flight, after watching a few calls.
05
Recruiter trust
Trust needed an escape hatch

Shortlists stayed at 4 to 6%. Recruiters verified everything.

Manual overrideReassessmentRejection reasonsProof of attempts
↳ We built for verification instead of fighting it.
🎧
Multi-call interview timeline
Screen pending export: per-call activity and recordings, call 1 inconclusive, call 2 completed.
Phase 1 validated the product and taught us where it strained. What remained unsolved became Phase 2.

Candidates Onboarding Flow Revamp

Highest qualification
Personal details
Profile ready

Overview

Onboarding was built as a single linear data collection flow applied uniformly to every user, whatever role they came for. A delivery partner and a software developer complete the same number of steps, at the same depth, before either one reaches the job feed. It converts at 67%, takes ~12 minutes, and asks for 6 to 12 steps of input before delivering any value.

The fix is role based branching plus progressive enrichment: split onboarding into a Fast Track Flow for gig roles and a Deep Profile Flow for regular roles, and recover the deferred data later, at moments when the user is already motivated.

My role
I led this project end to end.
Timeline
2 months
Responsibilities
Primary research, feature prioritisation, UI and interaction design, usability testing
Team
Stakeholder team, Product Manager, Dev team

01 · Problem statement

One long flow for everyone, whatever job they came for.

Onboarding asked every user the same questions, in the same order, to the same depth. A delivery partner looking for gig work and a software developer complete the same number of steps before either of them reaches the job feed — even though what we need to know to match them could hardly be more different. One needs a licence, a vehicle and an area. The other needs a title, a stack and a salary band. We asked both for all of it.

That flow runs 6 to 12 steps and about twelve minutes, and none of it returns anything on the way: no job, no salary, no sign that employers are hiring. Applying the deepest profile we might ever need to every user, regardless of role, is what pushes people out. A third of everyone who logs in never finishes.

It did branch, but only on education level and work status. It never branched on the thing that decides how much we actually need to know: the job the person is doing today.

02 · Who our users are

A third of the people we acquire come in for gig work.

The flow they landed in was built around credentials, which is not what a gig job needs. The sharpest version of that mismatch sits inside this third.

31%of all acquisitions come in for gig roles

Delivery, driving, logistics, security. A third of the base has low tolerance for form filling.

45%of gig users are “experienced”

Which routes almost half of them into the longest experience subflow we have.

~10%are gig seeking and graduate or above

The sharpest wedge. They want a gig job but get routed through the deepest, most credential heavy flow, because we branch on education instead of role.

Insight

A graduate applying for a delivery role is treated as a graduate first and a gig seeker second. They absorb 10 to 12 steps of profiling for a job that needs almost none of it.

03 · Objectives

Two objectives, one of them numeric.

Objective 01 · Onboarding conversion ~80% ▲ from 67% today Baseline 67% · Target ~80%
Objective 02 · UI revamp Clarity and perceived length

Fewer fields people misread, and a flow that feels shorter than the one it replaces. Read from dropoff per step, and from what users say in testing.

Baseline qualitative · Target qualitative

04 · The current flow

Everyone walked the same path.

After login, every user got the same screens in the same order. Only two answers changed anything: education level and work status. Those two splits made four different paths: the shortest was seven screens, the longest sixteen.

login → language → basic details → about me → location → education → experience → [fresher / experienced] → experience details → company & industry → job detail → skills → salary → role type → preferred language → preferred job role → resume → job feed
Everyone starts here
Language
Language4.7% dropped here80,134 reached it
Basic details · viewed
Basic details · viewed3.3% dropped here76,337 reached it
Basic details · cohort split
Basic details · cohort split2.6% dropped here36,460 reached it
About me
About me3.8% dropped here35,518 reached it
Location
Location4.4% dropped here34,421 reached it
Education
Education11.1% dropped here32,915 reached it
Experience
Experiencesplits into experienced and fresher29,273 reached it

How to read this: the percentage on a screen in the funnel sheet is how many people landed on it, so the shortfall belongs to the screen before. 95.3% land on basic details, which means 4.7% gave up on language. Only 90.5% reach language at all, so a further 9.5% leave between login and the first question.

Graduate / diploma → experienced: nine more screens
Years of experience
Years of experience2.1% dropped here12,706 reached it
Company & industry
Company & industry7.2% dropped here12,443 reached it
Job details
Job details9.4% dropped here11,547 reached it
Job role
Job role1.6% dropped here10,458 reached it
Skills
Skills0.9% dropped here10,292 reached it
Salary
Salary1.8% dropped here10,196 reached it
Role type
Role type1.5% dropped here10,012 reached it
Interest selected
Interest selected31.6% dropped here9,768 reached it
Resume upload
Resume uploadlast screen of this path6,783 reached it
Graduate / diploma → fresher
Skills
Skills1.6% dropped here16,609 reached it
Language
Language0.8% dropped here16,347 reached it
Role type
Role type22.0% dropped here16,214 reached it
Resume upload
Resume uploadlast screen of this path12,652 reached it
10th / 12th pass
Basic details · cohort split
Basic details · cohort split1.1% dropped here37,359 reached it
About me
About me4.3% dropped here36,952 reached it
Location
Location5.6% dropped here35,369 reached it
Experience
Experiencesplits into experienced and fresher33,396 reached it
10th / 12th pass → experienced
Years of experience
Years of experience7.6% dropped here17,450 reached it
Company & industry
Company & industry23.8% dropped here16,122 reached it
Job details
Job details7.6% dropped here12,281 reached it
Language
Language1.2% dropped here11,344 reached it
Role type
Role typelast screen of this path11,210 reached it
10th / 12th pass → fresher
Language
Language0.8% dropped here19,737 reached it
Role type
Role typelast screen of this path19,585 reached it

05 · Qualitative baseline

Four pain points from the initial interviews.

Pain point 01 “It’s too long.”

The most common and most direct complaint across the initial round.

Pain point 02 “I don’t know what to put here.”

Fields like Job Role and Industry. The taxonomy is ours, not theirs.

Pain point 03 “Why are you asking me this?”

Questions are irrelevant and not personalised. Gig workers get white collar questions.

Pain point 04 “I can’t tell what this screen wants.”

Visuals are not explanatory, and are actively confusing in places.

What this leads to: 33% funnel dropoff, and widespread incorrect data in role, industry and similar taxonomy driven fields.

06 · What was going wrong in the data

Two thirds of people who log in finish onboarding. The third we lose is not random.

10%39%42%66%40%39%30%50%88%95%96%95% 7%14%14%12%20% Login success11,561Resume path1,167Graduate4,482Non graduate4,885Onboarding success766Experienced1,798Fresher1,750Experienced1,460Fresher2,448Onboarding success1,579Onboarding success1,669Onboarding success1,402Onboarding success2,330 Onboarding success, all paths7,746 · 67%

Roughly 3,800 people a cycle log in and never finish. The resume path and the graduate experienced branch are the two weakest routes in the product.

Then I cross referenced completion against how many steps each segment clears.

Non graduate fresher6 to 7 steps95%
Graduate fresher7 to 8 steps95%
Non graduate experienced9 to 10 steps~96%
Graduate experienced10 to 12 steps88%

Where people actually dropped

How to read the funnel: the figure on a screen is the share of the previous screen’s users who arrived. So the loss belongs to the screen before it: 88.94% reaching the work status screen means 11% gave up on the education screen. Below are the screens people were actually standing on when they left.

In flow order: the screens people were on when they gave up
Education
Education11.1% dropped herenever reached work status · everyone
Company & industry
Company & industry23.8% dropped herenever reached job details · 10th / 12th · experienced
Job details
Job details9.4% dropped herenever reached job role · graduate · experienced
Resume upload
Resume upload22.0% dropped heregraduate · fresher

Company and industry is the single worst screen in the product; it loses nearly a quarter of the 10th/12th experienced branch. Education costs 11% of everybody. And barely anyone gets as far as the resume screen: 22% of graduate freshers and 31.6% of graduate experienced users drop on the step immediately before it. Every one of these is typing heavy; the screens people tapped through held above 97%.

What the data was telling us

The longer the flow, the more people quit.

People weren’t dropping off because of who they were. They dropped off because of how far we made them walk.

07 · Defining the problem

The fix had to be structural, not cosmetic.

Better copy and better buttons don’t help a flow that is simply too long for the person walking it. The path itself had to get shorter, and it had to get shorter for the people who need it shorter, without flattening it for the people who don’t.

The usual objection is “we’ll lose match quality.” But users were already filling role and industry wrongly because they didn’t understand what was being asked. We were already losing data quality. Collecting fewer fields people actually understand may produce a better database than collecting many they guess at.

1 · Value first

Show jobs as early as possible. Value precedes data.

2 · Effort matching

Input required should be proportional to intent and role complexity. A delivery role should not cost the same as a developer role.

3 · Progressive profiling

Collect only what is essential upfront. Defer the rest to a moment where the user has a reason to give it.

08 · How we decided to work

In phases, so the data could tell us what actually worked.

The tempting move was to redraw the whole flow in one release. We deliberately didn’t. Change fifteen screens at once and you get one number and no explanation: if completion moves, you have no idea which change earned it, and no way to know you left points on the table.

Phase 1 · Milestone 1

Only the flow up to education

Three test variants on a single question: how much education do we need upfront? Nothing after education is touched.

Phase 2 · Milestone 2

The rest of the flow

Built on whichever variant wins. Persona branching, and the experience section reworked per persona.

Then · field research

Talk to real users

After the two rounds of measured change, sit with the people the flow is for, and find the problems the funnel could never explain.

Phase 3 · Milestone 3

Revamped UX, shorter flow

Scope set by the findings rather than by inference.

Each phase inherits the previous one’s answer instead of relitigating it. Onboarding is also the only gate into the product, so a regression here suppresses every downstream metric at once, another reason to move in measurable steps.

09 · Phase 1 · Milestone 1

One variable, three levels: how much education do we need upfront?

Scope was cut deliberately: the language screen removed, basic details and location reworked, and nothing after education touched. All three variants carry those same front half changes, so any difference between them is down to education handling alone.

Test 1: Baseline+low

Education stays where it is

No structural change. Carries only the front half improvements; everything after Location is identical to production. Isolates the value of those changes, and acts as control for the other two.

Test 2: Repositionmedium

Education moves to the last step

Tests whether the dropoff follows the screen or is caused by its position. It distinguishes “education is hard” from “education is hard this early.”

Test 3: Deferhigh

Education moves last and becomes skippable

Education sits at the final step of onboarding and can be skipped outright, then asked later when the user applies to a job and the field’s purpose is self evident. The most aggressive test, and a dress rehearsal for the enrichment model Milestone 2 depends on.

Test 3 is the cheap proof. If deferring education to job apply works without collapsing data completeness, the whole progressive profiling thesis gains evidence for the price of one variant.

Test 2 is the safety net. If Test 3 wins on completion but craters education capture, Test 2 offers most of the gain at much lower data risk.

What “ask for it later” actually looks like

In Test 3 the education screen is skippable during onboarding. The fields come back inside the job application, bundled with the document checks that job already requires. By then the candidate has found something they want, so the question has an obvious reason behind it.

The assessment flow: education asked during job application
Job feed
Job feedThe candidate finds a job they want. Nothing has been asked of them yet.
Job details
Job detailsThey tap Apply for job.
Assessment: step 1 of 2
Assessment: step 1 of 2“Following items are mandatory for this job.” The document checks that this job actually requires: Aadhaar, PAN, vehicle RC, driving licence.
Education: step 2 of 2
Education: step 2 of 2The fields skipped during onboarding, asked here instead. “Education info will be saved in your profile.”
Education, filled
Education, filledDegree, completion year, specialisation, college, school medium, then Finish Application.
Application submitted
Application submittedGood match, job applied, AI screening pending. The profile is complete and the candidate has applied.

How we planned to measure it

What we watchedHow many people finished the flow, which screen they left on, and how long it took them to see a first job.
What must not breakWe still need education data. We check how much of it we have after 24 hours and after 7 days. In Test 3 nobody is asked for it during signup, so a finished signup counts for nothing if the profile stays empty.
The real testDo these people apply for a job on day one? A faster signup is only a win if it produces applicants, not just finished forms.
How we split the dataGig work against regular jobs. Graduates against everyone else. Freshers against experienced. And the graduate experienced group on its own, because that is our longest path.
Results

All three variants ran for two weeks against a held back control. The completion read is in the next section; education capture and day one apply rate are still to come.

Designs: Figma: experiment milestone 1 & 2

The design was shipped after minor back and forth with engineering and PMs

As usual !! :)

NOW COMES THE FUN PART

DID USERS ACTUALLY USE IT?

Milestone 1 · Results

Position wasn’t the problem. The requirement was.

Every variant carried the same front half changes, so Test 1 is the honest control for the other two: whatever separates Test 1 from Test 2 and Test 3 is education handling alone. Read that way, the result is sharper than the headline number suggests.

ArmWhat changedCompleted
Control Production flow, untouched 78.74%
Test 1 · Baseline+ Front half changes only; education left where it is 79.65%
Test 2 · Reposition Education moved to the last step, still required 79.65%
Test 3 · Defer Education moved last and made skippable, asked again at job apply 80.55%

Completion is login success through to interest selected. 94,710 logins over two weeks, roughly 23,600 per arm, with traffic split evenly across the four arms.

What each change was actually worth

The front half rework

Language screen removed, basic details and location reworked. Control to Test 1.

Moving education last

Same screen, same fields, final position. Test 1 to Test 2. Nothing moved.

Letting people skip it

Identical position to Test 2. The only difference is that answering became optional.

10 · Phase 2 · Milestone 2

Now the second half, and the branch that makes it shorter.

Milestone 2 picks up everything after the work status screen, on top of the Milestone 1 changes: the front of the flow reworked to location, education moved to the end. This is where the ~20% drop across the experience section lives.

Where that 20% actually goes

ScreenDrop
Years of experience2%
Company & industry name9%
Job details10%
Job role1.5%
Skills~1%
Salary2%
Preferred role~1%

Two screens carry almost all of it: company and industry, and job details.

How we tell gig from non gig

We do not ask for a job title on the first screen. The classification happens later, from a question we were already asking.

The two screens the routing runs on
Work status
Work statusConfirm your work status. Selecting “I’m working / I have work experience” opens the experience block.
Years of experience + title
Years of experience + titleThe next screen: total years of experience and current / latest job title. This title is the routing signal.
Where the classification happens
  • Confirm your work status → “I’m working / I have work experience”
  • Next screen: total years of experience + current / latest job title
Is that title one of the gig roles?
Yes →

Gig flow: shorter

Delivery · Driver · Logistics · Cook · Security. The experience detail screen becomes optional or disappears entirely, and no notice period is asked of anyone already employed.

No →

Non gig flow: standard depth

Everything else. The detailed work history stays, but industry is gone and salary is asked monthly.

The gig list is configurable, and it is the most load bearing definition in the project; every branching decision and measurement cut depends on it. For the initial rollout it is Delivery, Driver, Logistics, Cook and Security. It needs to hold the titles people actually type, including vernacular and spelling variants.

Two changes that apply to every variant

Industry selection is removed. It was a 9% drop on a field already being derived from the company name.

Salary is asked monthly, not as annual CTC. Someone paid per delivery or per shift can answer monthly without doing arithmetic first.

The flow each journey walks

Basic details
  • Highest level of education
  • Resume / LinkedIn link (diploma & above)
  • Full name
  • Date of birth
  • Gender
  • Email address
Location
  • Current city
  • Residential area
Work status
Working, or student / intern. This answer opens the branch
Experienced user: GIGeducation last, no resume
Experience
  • Total years of experience
  • Current job title
The routing signal
Current company details
  • Company name
  • Start date
  • Currently working in same company
  • End date
All optional, skippable
Job role
Suggested from the title
Skills
  • Skills
  • Monthly salary
Language
  • English proficiency
  • Languages you can speak
Preferences
  • Preferred job role
  • Preferred location
Education
  • Highest level of education
  • College name
  • Degree
  • Specialisation
  • Completed year
Experienced user: NON GIGfull depth
Experience
  • Total years of experience
Current company details
  • Current job title
  • Company name
  • Start date
  • Currently working in same company
  • End date
Job role
Suggested from the title
Skills
  • Skills
  • Monthly salary
Language
  • English proficiency
  • Languages you can speak
Education
  • Highest level of education
  • College name
  • Degree
  • Specialisation
  • Completed year
Preferences
  • Preferred job role
  • Preferred location
Resume
Upload or skip
Fresher / student user
Education
  • Highest level of education
  • College name
  • Degree
  • Specialisation
  • Completed year
Skills
Prefilled, editable
Language
  • English proficiency
  • Languages you can speak
Preferences
  • Preferred job role
  • Preferred location
Resume
Upload or skip
10th / 12th passouts: with experience
Experience
  • Total years of experience
Current company details
  • Current job title
  • Company name
  • Start date
  • Currently working in same company
  • End date
Job role
Suggested from the title
Skills
  • Skills
  • Monthly salary
Language
  • English proficiency
  • Languages you can speak
Preferences
  • Preferred job role
  • Preferred location
10th / 12th passouts: without experiencethe shortest path
Skills
Prefilled, editable
Language
  • English proficiency
  • Languages you can speak
Preferences
  • Preferred job role
  • Preferred location

Three shared screens, then the split. Scroll a lane sideways to see it all. A gig driver clears four groups; a non gig graduate clears five plus a resume.

The experiment: one control, two variants

GigNon gig
ControlFlow as it is todayFlow as it is today
Test 1Experience details skippable; no notice period if employed Experience details skippable; anything skipped is asked in the assessment flow
Test 2No experience screen at all, captured in the assessment flow insteadExperience details mandatory

Allocation is by user ID: Test 1 on IDs ending 00 to 29, Test 2 on 30 to 59, control on 60 to 99. Variant 2 also asks for expected monthly salary rather than current.

The reworked screens

Milestone 2: the second half as designed
Years of experience + title
Years of experience + titleOpens with the cheapest question: total years, then the title.
Current job details
Current job detailsCompany and dates. Skippable in variant 1, gone for gig in variant 2.
Roles and responsibilities
Roles and responsibilitiesRole arrives as suggestions from the title. Recognition, not recall.
Skills + monthly salary
Skills + monthly salarySkills prefilled. Salary monthly, with the reason stated.
English + languages
English + languagesEnglish and languages collapsed into one light step.
Job preferences
Job preferencesPreferred role and location.
Results

The first read on these variants is in the next section. Step level dropoff, time to first job view and D1 apply rate are still to come.

Milestone 2 · Results

The gig branch is where the shortening actually paid.

Milestone 2 tested the experience block: remove those screens for gig roles, or leave them in but let people skip and answer later. Both beat control, and the effect is far larger inside the gig cut than across everyone — which is what the branching was built to do.

CutWhat was testedResult
Gig users Experience screens removed from the flow +10.5 pp
Gig users Experience screens kept, but skippable +7.2 pp
Everyone Skip during onboarding, asked again after apply 76%

Profile completion against control. The all users row is 76% against a 70% control, so +6 pp. Removing the screens outright beats making them optional by roughly 3 pp inside the gig cut.

Read this apart from Milestone 1

The baselines are not the same number. Milestone 1 measured a 78.74% control on login to interest selected. This reads a 70% control on profile completion. Different gate, different population, so the two gains cannot be added together or compared directly — each one only means something against its own control.

What Milestone 1 established still holds. There, moving a required screen changed nothing and making it optional moved everything. Here the same shape appears again, one step further in: the biggest gain comes from removing the ask entirely for people it was never serving, not from repositioning it.

What we are changing, and what we are not changing yet

Not shipping on this aloneNo product changes go in off these numbers by themselves. Microsoft Clarity went in on 17 September, and the session recordings come first, to confirm people are behaving on these screens the way the funnel implies.
Non gig flowsThe non gig branch has the weaker result and the clearer headroom. Those changes get finalised off the Clarity read rather than guessed at now.
The gig definitionDelivery, driver, logistics, cook and security is a narrow list. Broadening it and re validating on a larger base is the next test, since every branching decision in the project hangs on that definition.
The deferred questionsAsking the skipped questions after apply does lose people. The completion win is real, but it moves the drop rather than removing it, and where it lands now needs its own measurement.
Not yet measured

Sample size and significance were not reported with this read, so treat the three figures as directional until the cut sizes are attached. Step level dropoff, time to first job view and D1 apply rate are still outstanding for this milestone, as is the size of the post apply drop on the deferred questions.

11 · Then we talked to users

The data told us where people left. It could never tell us why.

Two rounds of measured change in, we still hadn’t sat with a real user since the original pain point interviews. Seven moderated sessions, in person, think aloud, on the interactive prototype. The first field research on this flow in a long time.

User research

A moderated usability study on the new flow.

Seven participants, one at a time, on a clickable prototype of the shortened flow. Each was asked to sign up as themselves while thinking aloud, and we recorded where they hesitated, what they misread, and the screen they stopped on. The question was never whether people liked the design. It was whether a real jobseeker can get to the end of it without help.

Two decisions about how we ran the sessions did a disproportionate amount of the work, because each one exposed a failure the obvious setup would have hidden.

Prototype in English, sessions in Hindi

We moderated in Hindi but left the prototype in English, exactly as it ships. Translating it for the test would have been the comfortable choice, and it would have concealed the single most severe finding in the study: a participant who could not read the flow at all, and so never started it. Testing the real language is what made that visible.

Word recall after the task, not during

We asked people to define CTC, Specialisation and School medium only after they had finished, without showing the screens again. Mid-task, someone who guesses and moves on looks identical to someone who understands — both just tap Next. Asking afterwards, from memory, is what separates comprehension from a lucky guess.

We explicitly did not test visual polish. Aesthetic feedback was redirected to function: “would that change whether you could finish signing up?” That is why every finding below is structural.

Who we sat with

#ProfileLanguageOutcome
1Delivery agent, college dropoutReads Hindi + English, limited spokenReached final screen, ~11 min
2Delivery agent, graduateFluent bothDropped off: current company details
3Security supervisor + deliveryFluent bothStopping point not recorded
4Security personnelSpeaks Hindi · cannot read Hindi or EnglishCould not start
5Delivery agentBasic bothDropped off: current job details
6Sales executive, 3 yrs (contrast case)Fluent bothCompleted smoothly: no confusion
712th pass, no experience (fresher path)Not recordedCompleted easily: few minutes

Two clean completions: one white collar, one on the short fresher path. One reached the end with heavy friction, two dropped at the same screen, one couldn’t start. Every clean completion came from either a white collar profile or a shortened path.

12 · What we found

Twelve findings, grouped by how much they cost us.

12findings
5critical
4high
3medium & below
7participants
CriticalBlocks completion, or invalidates something we had already decided
011 of 7 · the contrast case

The flow isn’t broken. It’s mismatched to the segment.

One participant (graduate, sales executive, fluent, tech savvy) completed the whole flow with zero confusion. Every critical and high finding in the study came from the six blue collar participants.

So what

It validates the branching thesis from the qualitative side, and tells us the Deep Profile flow should be protected, not redesigned. Evidence for what to leave alone is rarer than evidence for what to change.

021 of 7 · could not start

For some users this isn’t friction. It’s a hard blocker.

One participant speaks Hindi fluently but cannot read Hindi or English. He could not start at all. He is not technology averse; he uses a gated community app competently. The barrier is literacy.

So what

A different category of problem: everyone else struggled through, he couldn’t begin. It’s also upstream of the router: he can’t enter a job title, so he can’t be classified at all. Our architecture has no answer for him.

033 complained · 1 clean on short path

Shortening the flow is a confirmed fix, not just a complaint.

The participant on the fresher path, which skips Experience Info and Skills entirely, finished easily in a few minutes. Three others complained, unprompted, that the flow was too long.

So what

The first direct causal evidence that length itself is the problem rather than field wording. A shorter path produced a clean completion in the same segment where longer paths produced dropoffs.

Job details
04near universal

Custom job title entry is broken for nearly everyone.

Entering a title that isn’t in the suggested list failed in almost every session, regardless of literacy, education or English comfort. That independence points at the interaction, not the audience.

So what

This is the finding that most threatens our own architecture. The router classifies on current job title, and the most broken interaction in the product is the router’s own input. So it is a Milestone 2 blocker, not an M3 line item, and it has to be rebuilt as guided, low typing selection.

Education
053 of 7

The education section assumes a graduate path.

Three participants hesitated or lost interest here. A college dropout was still asked for college details. One holding both 12th pass and a diploma couldn’t tell which applied, and that step took him longer than anything else. Another’s degree was missing from the list entirely.

So what

This independently validates the Milestone 1 hypothesis from a different evidence base. We inferred education was the bottleneck from the funnel; users confirmed it and said why: the schema doesn’t describe them. That favours Test 3 or a conditional form.

HighCosts users or costs data quality
Company & industry
062 dropped · 1 confused

“Current company details” is a repeated, independent dropoff.

Two participants abandoned at the identical step, independently. A third was confused but continued. Gig work often has no single traditional employer, so the question has no clean answer.

So what

Two independent dropoffs at one screen is a signal, not a coincidence, and it needs a gig shaped version of work history, not a shortened version of the white collar one.

Job role
073 satisficed · 2 inaccurate

Users satisfice, and that is a data quality problem.

When a field took effort, people picked the first suggestion or entered something rough. Three defaulted to the first suggestion on title or skills. Two entered inaccurate values: a random salary, and a company name that was simply made up.

So what

Empirical confirmation that we are already losing data quality. Profiles built by satisficing underrepresent what these users can do, which degrades matching, which produces fewer calls.

082 of 7 · existing users

The value exchange is already in deficit.

Two participants were existing users, already frustrated at getting no calls, and said so unprompted. One had found his current job through a competitor instead.

“We are asking so many fields but still we are not getting any calls.”
So what

Every field is being judged against existing disappointment, not a blank slate. This validates Value First but also bounds what onboarding can fix: a shorter form will not repair a broken value exchange; it just gets users to the disappointment faster.

Experience
091 explicit · medium / high

The fresher / experienced binary doesn’t fit gig work.

Occasional gig work is not a regular job, but it isn’t inexperience either. The same forced single choice appeared at qualification for someone holding both 12th pass and a diploma.

So what

Our funnel branches on this binary and Milestone 2 layers persona branching on top of it. If the categories don’t match how users see themselves, we are branching on noise, and it may explain the unattributed users who exit at work status.

Medium & belowWorth fixing, not worth blocking on
Location
104 of 5 who reached it

Zero successful adoption of the preferred city picker.

Four of the five whose location behaviour was recorded kept their current location, tapping “Yes” without reading, then adding the same city again. One read it and deliberately declined. Nobody used the picker as designed.

So what

A screen with effectively zero successful engagement doesn’t justify its prominence. Two dead ends surfaced here too: a missing residential area, and no path forward at all when location permission was denied.

About me
113 of 7

Email costs disproportionate time, even when optional.

Three spent longer on the email field than on almost anything else. One filled it carefully despite it already being optional; he never noticed it was skippable, and was disengaged everywhere else.

So what

Optional isn’t optional if it doesn’t look optional. Cheap fix, meaningful time saving.

Resume upload
121 of 7 · low / medium

The progress bar and resume nudge go unnoticed.

The progress bar didn’t appear to inform pacing or motivation, and the resume nudge was ignored entirely.

So what

Neither is earning its screen space in this segment.

13 · The verdict on our own plan

Read the research as a judgment on decisions we had already made.

FindingEffect on the plan
ConfirmedSegmentation is the right frameValidates the whole branching thesis: proceed with confidence
ConfirmedEducation is a genuine bottleneckValidates Milestone 1; favours Test 3 or a conditional form
ConfirmedLength itself is the problemValidates Fast Track’s ≤2 step goal
ConfirmedWe already have a data quality problemRemoves the main objection to deferral
ChangedThe router’s own input is the most broken interactionPromotes custom title redesign from M3 to an M2 blocker
ChangedLiteracy is a hard blocker upstream of routingNew workstream: the architecture has no answer for it
ChangedDon’t redesign the non gig flowDeep Profile is a preservation job, not a redesign job
BoundedOnboarding can’t fix the value exchangeSets a limit on what any milestone can claim

14 · Phase 3 · Milestone 3

What the sessions told us to build.

Every item below started as something we watched a person struggle with, and is described the way they described it. Nothing here is inferred from the funnel alone.

Two things have to be fixed before the split can ship

Let people describe their job in their own words

Almost nobody could enter a job title that was not already in our suggestion list, whatever their education or English. That same answer is what decides which flow a person gets, so when it fails we either send them down the wrong path or stop them entirely. The branching cannot ship on top of a broken question.

Give people who cannot read the flow a way through it

One man speaks Hindi fluently but cannot read Hindi or English. He could not get started at all, and he is not new to phones: he already uses his gated community app competently. He never reaches either flow, because the first thing we ask him to do is read and type. We have no answer for him yet, so this became its own piece of work rather than a line in this milestone.

Then the flow itself, hardest problems first

01Ask about education only when it appliesThree of seven stalled here. A college dropout was still asked for college details. One man held both 12th pass and a diploma and could not tell which single option to choose. Another found his degree was not on the list at all.Critical
02Ask about work the way gig work actually happensTwo people quit at the same screen, current company details. Gig work often has no single employer and no clean start and end date, so the question has no honest answer.Critical
03Ask for less, and say why we are askingWhen a field took effort people entered whatever moved them forward. Three took the first suggestion offered. Two gave a salary and a company name that were not real.Critical
04Use the words people already useCTC, Specialisation, School medium and AI suggested skills were answered without being understood. Asking people what those words meant after they finished showed they never knew.Critical
05Give every list a way outDegrees and residential areas were missing from our lists, and there was no “other, please specify” to fall back on. A missing option is a dead end, not an inconvenience.High
06Rebuild the fresher or experienced questionGig work fits neither answer. Occasional gig work is not a regular job, but it is not inexperience either. We branch the whole flow on this answer.High
07Catch answers that are obviously not realA random salary and an invented company name both went straight through. Every one of those degrades the matching that is supposed to get the person a call.High
08Collect the rest later, when there is a reason to give itThe same fields asked after someone applies for a job have an obvious purpose, and people are already motivated. Asked upfront, they are just a toll.High
09Make optional fields look optionalThree spent longer on email than on almost anything else. One filled it in carefully and never noticed he was allowed to skip it.Medium
10Fix the location screenFour of the five whose location we recorded kept their current city, then typed the same city again as their preferred one. One had no way forward at all after refusing location permission.Medium
11Progress bar and resume nudgeNobody noticed either of them. They either need to do something useful or give the space back.Low

The shortened flow, as built

Screens exported from the Milestone 3 prototype, the same build the sessions above were run on.

Milestone 3: the revamped flow
Highest qualification
Highest qualificationThe flow opens on one question, under a two minute promise. Six levels including ITI and Diploma, and it asks for the highest one, so the man who held both 12th pass and a diploma no longer has to decide which of the two he is.
Personal details, or a resume
Personal details, or a resumeAuto fill from a resume sits above the manual form rather than after it. Name, gender, date of birth and email in a single screen.
Where do you live
Where do you liveOne city question with a use my current location shortcut. The old flow asked current city and preferred city separately, and four of the five people whose location we recorded typed the same city twice.
Work status
Work statusThe branch point. Working opens the experience block; fresher, student or intern takes the short path.
Roles and responsibilities
Roles and responsibilitiesRole is entered by search plus suggestions instead of a fixed list, so a job title that is not already in our data can still be described. This was the question that almost nobody could get past.
Skills and salary
Skills and salarySkills carry over from the role. Salary is asked as a monthly figure rather than CTC, and the screen states why it is being asked.
Preferred roles
Preferred rolesPreferred roles, then an optional question about working outside the current city, with preferred locations added only if the answer is yes.
Language
LanguageEnglish level and any other languages, collapsed into one light screen.
Documents and assets
Documents and assetsA new screen for frontline roles: which licences and vehicles the person has, with the reason for asking stated in the header.
Education, last and skippable
Education, last and skippableEducation moves to the end of the flow and carries a Skip. This is the Test 3 deferral built in rather than assumed.
Two paths through it
Resume first
Resume firstThe same entry screen with the resume path taken. Everything the resume can fill, it fills.
Fresher path
Fresher pathSelecting fresher, student or intern skips the whole experience block.

What we are deliberately not touching: the longer flow for regular jobs. The one participant it was built for, a graduate sales executive, finished it without a single moment of confusion. We have evidence that it works and no evidence that it needs changing, so we leave it alone.

What onboarding cannot fix: two people told us, without being asked, that they fill everything in and still get no calls. One had found his job through a competitor. A shorter form does not repair that, and saying so now matters: if applications do not go up after this, onboarding may not be the reason.