Moov 37
the agentic job search experience
Saas
Product Strategist
Product designer
UX Researcher
Interaction Designer
Ai in hr
Recruitment technology
Imagine a world where discovering the perfect candidate and landing your ideal job, is as effortless as a conversation.
About Moov 37
MOOV 37 IS A NEXT GEN AGENTIC HIRING PLATFORM DESIGNED TO REVOLUTIONIZE hiring for recruiters and job seekers. the app leverages ai agents that reason, act and adapt aiming to radically simplify the process and making it intelligent, interactive, and real-time.
a glimpse of the ai agent powered—moov 37's landing page.
The platform was conceptualized for a Y-combinator pitch and built as a working prototype. While the current live version is a minimal v1 without ai agents due to budget constraints, this case study focuses on the agentic-first design vision that powered the prototype.
High-level Problem
traditional platforms (linkedin, naukri, uplers) suffer from manual, fragmented workflows. the global market is sluggish and competitive—especially for international students and early-career professionals.
Job seeker problem
/ form fatigue from entering repititive data across platforms.
/ long onboarding processes with low reuse of resume content.
/ unclear application flows and little match-based guidance.
Recruiter problem
/ time-consuming setup before receiving useful candidate matches.
/ over-reliance on filters that often produce low-quality results.
/ difficulty generating compelling job posts efficiently.
Opportunity
Reimagine the hiring process by simplifying onboarding, reducing form clutter, and enabling AI-assisted, guided workflows on both sides.
Deep-Dive
Into The Problem
to understand current gaps in the hiring journey, I conducted a competitive analysis of leading job platforms like uplers, eightfold, and linkedin recruiter, Identifying pain points across both candidate and recruiter flows.
according to a 2023 glassdoor study, candidates abandon 60% of applications mid way due to long or repetitive forms. Meanwhile, recruiters report spending nearly 30% of their time filtering irrelevant profiles.
Candidate-side
/ long, repititive forms as they fill in the same information across multiple platforms.
/ while job portals expect them to "Upload resume" and also fill in the same details again, most platforms do not parse based on uploaded content.
/ no clear feedback on how well they matched job roles.
/ little sense of progress or personalisation during onboarding.
Recruiter-side
/ high time spent on setup before seeing meaningful candidate results.
/ Quality of applicants is often poor unless filters are used strategically.
/ No time or expertise to write strong job posts that attract the right talent.
/ recruiters are starting to experiment with chat-GPT to write job posts or review profiles which shows a behavioural shift.
️TL;DR gist of my research
candidates are tired of repeatedly filling out the same data across multiple platforms, facing long forms and poor matching experiences.
recruiters, on the other hand spend time manually entering jobs and navigating low quality filters.
there is a need to reduce redundancy, guide users and make hiring feel less like paperwork and more like progress.
Solution Hypothesis
Approach 1: dashboard-first approach inspired by uplers
wireframe: a control center like dashboard with stats.
What
candidates and recruiters view an application similar to a control center. Users would see real-time stats (eg.: Number of views, status updates, recruiter response) and manage everything from a centralised dashboard.
Pros
/ empowers recruiters to quickly scan, filter, and follow up without digging into systems.
/ encourages accountability both sides with better visibility.
/ works well with power users managing multiple jobs or candidates.
Cons
/ dashboard with multiple things showing up at the same time can become overwhelming.
/ doesn't fully solve for the friction of form filling or discovery.
/ could mirror existing ATS dashboards if not differentiated visually and functionally.
Approach 2: form-like structure
with contextual suggestions
Approach 2: form-like structure with contextual suggestions
wireframe: a control center like dashboard with stats.
What
the idea leaned into making form-filling as painless as possible, no fluff, but heavy ux optimisation. it relies on inputs and clever contextual suggestions. it takes inspiration from well-designed onboarding flows like notion or superhuman.
Pros
/ familiar ui pattern that requires no learning curve.
/ low implementation effort with maximum usability impact.
/ easier to validate future ux improvements.
Cons
/ does not differentiate from competitors, might feel too normal.
/ makes it look like—"just another job application".
/ Lacks the excitement or "wow factor" for an early stage pitch/vision.
/ does not tap into the potential of future proof patterns.
Approach 3: ai Agent-powered with
natural language and fallback ui
Approach 3: ai Agent-powered with natural language and fallback ui
wireframe: a control center like dashboard with stats.
What
the idea focussed on reducing form fatigue by letting users describe their needs conversationally—"I'm looking for hybrid tech roles in mumbai under 10lpa", and allowing the agent to handle the rest. the agent could also assist recruiters and candidates with setting up roles and interests and reviewing matched jobs or candidates.
Pros
/ drastically reduces manual input and makes the process feel fluid.
/ aligns with rising user expectations from tools like chat gpt, gemini, etc.
/ can adapt to both recruiters and candidates, creating a unified experience.
/ highly scalable with the right backend logic and model integrations.
Cons
/ high implementation cost; required real-time nlp, training and logic.
/ might create ambuiguity if guardrails or fallback ui isn't tight.
/ needs thoughtful handling of edge cases, especially for recruiter-side complexity.
/ can confuse users expecting traditional form-based interface.
Explorations—
Agentic architecture
Concern 1: long job forms discouraged completion
We aligned with an ai-powered chat interface to replace the long form. Information would be collected gradually through conversation, making the process feel lighter and more personal. the ai could also adapt questions based on earlier responses, helping users focus only on relevant details.
Pros
/ reduces intimidation from large forms.
/ feels more personal and interactive.
/ can adapt over time as ai learns.
Cons
/ can be slower for users who have all information ready.
/ interface might feel unfamiliar for users used to forms.
/ higher build cost than simple forms.
Concern 2: users felt lost when key
details scrolled out of view in chat
Concern 2: users felt lost when key details scrolled out of view in chat
We decided on a split-screen layout where the left 70% is the chat interface, and the right 30% is a persistent profile or job-posting/ profile panel. as the user chats, the ai Updates this panel in real-time so that information doesn't get lost in scroll. recruiters and candidates can directly edit any section by calling out that section in the chat.
Pros
/ prevents info from feeling buried in conversation.
/ gives users an always visible progress snapshot.
/ makes it easier to correct mistakes in the new ai job portal.
Cons
/ takes up screen space, reducing chat area.
/ requires careful ui balance so panel does not feel cramped.
Concern 2: users felt lost when key
details scrolled out of view in chat
Concern 3: recruiters don't have time to filter candidates
we introduced a match percentage with supporting reasons. the ai displays a score (e.g., 78% match) alongside key factors influencing that score (such as location fit, relevant experience). this would help recruiter and candidates quickly gauge suitability without looking here and there.
Pros
/ provides instant clarity on fit.
/ saves time by focussing only on high potential matches.
/ builds trust by showing reasoning.
Cons
/ match accuracy depends on quality of data input.
/ may give a false sense of certainty if reason context is incomplete.
Outcome
and MVP
we had ambitious plans for an agent-driven assistant, but for MVP we had to balance the vision with pragmatism. the shipped product was a structured, form-like interface, not glamorous, but a deliberate step to validate adoption and workflows before investing in agentic complexity.
this is What shipped as mvp.
this is where were heading.
Key outcomes
/ Recruiters published posting faster, from ~30 minutes to <10 minutes.
/ candidates often abandoned long registration portals. the mvp broke onboarding into 3 progressive steps, cutting major effort.
/ every input in the form was designed to "feed the agent" (skills, roles, preferences) that an agent ui in v2 could later query and edit in real-time.
/ we aligned with the client who needed something shippable and testable for market entry.
/ by choosing the simpler mvp, we de-risked launch timelines while keeping the door open for high-impact agentic features in v2.
Why this mattered
on the surface, yes—the mvp looks like just another form-based portal. but the value was in what it enabled:
we needed to prove that recruiters actually got better matches and candidates finished profiles fully—two things many portals fail at.
the form gave us structured, reliable data—the raw material the agentic system we envisioned for v2 would need.
jumping straight to an ai-first experience would've been costly and risky if the basics weren't solving the problem. the form de-risked the investment.
Reflections
designing moov 37's agent-powered interface forced me to abandon my usual "screen-by-screen" mindset. instead, i had to design for a living system that responded to user intent, kept context and adapted in real time.
Adaptive over static thinking
Our early dashboard-like concept fell short because it locked users into fixed flows. shifting to an agentic approach meant thinking in states—example, keeping the recruiter's job posting details pinned and editable white chat continued.
Progressive inputs beat long forms
Recruiters hated filling long job posting forms. we learned to let the agent collect details piece-by-piece, while showing how each input improved candidate match quality.
Explainability builds trust
early "match %" suggestions felt arbitrary. we improved them by showing 'why' a candidate matched, skills, location fit, education, etc—so recruiters didn't feel ai was guessing.
Trade-offs are inevitable
The agent was robust, but real estate was tight. We had to balance chat space with a persistent profile view, which meant cutting secondary features cluttered the UI.
I can address the elephant in your room
Let’s Connect
HRUDAY KULKARNI©