SAAS PLATFORM
AI-Powered Recruitment Platform to Boost SME Hiring Efficiency and Accuracy
Background
PolarisJobs, an Al start-up incubated by Harvard Innovation Labs, is on a mission to revolutionize the hiring process for small business and start-ups in US.
My deliverable
Conducted research, including interviews, and synthesizing insights.
Formulated strategic concepts, defined project scope.
Established a cohesive UI style guide for product consistency.
Developed components and interactive prototypes.
Impact
Our client successfully secured 200K pre-seed funding from Harvard Innovation Labs and Spark Grants.🎉
Award-Winning UXUI Design Concept.🏆
Effectively taking the product from concept and MVP to reality, the new product is launching in 2024.
Product Overview
AI-driven Job Posting: Type on Left, Generate on right, Enhance job posting speed by 15%.
Best Match Candidate: Multiple Fine-tune preferences for greater accuracy.
Easy Comparison: AI-driven intermediate analysis, Shorten shortlisting time by 10%.
AI Chatbot: 24/7 Customer Support and auto interview Scheduling, FAQ assistance.
Challenges of Target User
Recruitment Challenge for SMEs
55% of small businesses and start-ups DO NOT have a dedicated HR department.
74% of small businesses admit to hiring the wrong person for a role.
60% of small businesses report that hiring delays have a significant impact on their ability to operate effectively.
Start-ups owners spend 40% of their working hours on HR tasks that DO NOT generate income.
Hypothesis & Goals
Using AI-driven data analytics to improve hiring accuracy, efficiency and enhance professional HR support for small business in one Platform.
User Goals
To quickly find the right candidate with the help of AI-driven professional and professional HR assistance.
Business Goals
To provide a streamlined, all-in-one AI-assisted recruitment platform for small businesses, enhancing efficiency, fairness, and user trust throughout the hiring process.
Planning & project Set-up
Working with PM, I conducted client's brief to detailed project plan and scheme the right design process & method base on product scope, time and budget.
To understant the market
Existing AI recruitment platforms are content-heavy, lack streamlining due to a high learning curve.
Product Opportunity Gap
Competitive features include reducing AI bias, automating candidate outreach, empowering users to refine AI results, and obscuring certain demographic information to promote fairer recruitment practices.
What we MUST HAVE
Streamlined application tracking system and an automated screening system.
Painpoints from Interview
What do our users struggle/concern with the most?
Suspect AI Credibility
40% participant are sceptical about Al ability and ethicality, with the impression of AI gives random result with limited user control.
Unclear Job Description
56% participants do not have professional HR in their team and need external help to modify their formal job description. The inarticulate JD also cause high turnover rate.
Slow Shortlisting Process
68% start-ups frustrated the manual shortlisting process which resulted in hiring delays and impact on their ability to operate effectively.
Key Insights from painpoints
Design Principle review
We adopt a 'human first, AI-assist when necessary' approach with the most transparent and controllable AI intermediate tuning.
Let users trust AI
In our interviews, we discovered that many users hold reservations about AI in recruitment, viewing it as unethical to solely rely on AI scoring that may introduce bias. Therefore, the design team will prioritize building trust in AI while allowing sufficient manual oversight and adjustment capabilities.
Decision Making Methodology
To align user goals with business goals while considering constraints,
We defined feature deliverables for the MVP stage using Product Roadmap.
Affinity map & Ideation
Group brainstorming & Ideate from research insights, while keeping constraints in mind.
Prioritization
We scoped features based on research insights and got feedback from designers, engineers, and product managers. My task is to design on 2b onboarding, job description setting and shortlisting feature.
Narrow Down the problem for the MVP
How might we design an AI recruiting platform that auto writes precise job descriptions, effectively shortlists candidates, and minimizes AI bias to save users' time and costs?
Prioritization
We scoped features based on research insights and got feedback from designers, engineers, and product managers. My task is to design on 2b onboarding, job description setting and shortlisting feature.
User Flow Iteration
After testing the initial user flow with developers, I moved 'adjust hiring preference' before the shortlisting, and added pre-set question to establish hiring preference, it conserves 20% of AI computational power by reducing redundant calculations and searches, as the AI only needs to process candidate data once, based on the user-defined weights and preferences set beforehand.
Final MVP USer Flow
Key Layout Iterations
Behind the scene: some important decisions I made and why
More discussion & Testing!
These were the outcomes of design critique discussions with my managers & PMs, trial-and-errors, competitor analysis, and a lot of learnings of best practices. It was fun to see how the design evolved after discussion & testing!
Contingency Plan for edge case scenario in Shortlisting
To manage unexpected or exceptional situations, I designed this contingency plan for better manually shortlisting and in detail profile comparison, ensuring a seamless user experience even when AI scores don't go as planned.
As a designer, I made new technology more convincing & easy to use.
Multiple intermediate Tuning
Traditional resume scanning, which relies on a set of keywords, may lead to overlooking suitable candidates. Our AI model enables adjustments to the scoring system controlled by the recruiter's preferences. It identifies unstructured signal information and filters candidates from multiple dimensions, ensuring a more comprehensive evaluation.
1- AI Transparency: Explain where we get the data from and show the breakdown score from sources
2- Human First, AI-assist Second : Comprehensive scoring system
User Defined Scoring
AI Benchmark Scoring
3-User-Defined Weights and Control: Users can understand the scoring process, adjust scoring standards, and use benchmark weights, enhancing the accuracy and fairness of the final scores.
4- Leveraging Big Company Hiring Standards: AI-Driven Benchmarking uses leading companies' hiring standards to provide entrepreneur a professional hiring framework.
4- All the settings and backend algorithms provide the most accurate and personalized overall score.
Final Product Demo
Design Library
We use Material design system for MVP launching in limited time.⌛
Positive User Feedback
High user satisfaction ratings and positive reviews highlight the app's intuitive interface and powerful AI capabilities.
