Inspiration
HR teams use 6 different tools - Workday for employees, Greenhouse for hiring, Slack for surveys, Jira for performance. Data is scattered, so decisions are guesswork. We saw that 34% attrition happens without early warning because no tool connects these dots. We wanted to build one brain that connects all HR telemetry and gives AI-powered reasoning, not just charts. That's WorkMind AI.
What it does
WorkMind AI is a Workforce Intelligence Platform live at workmindai1.netlify.app
It does 6 things:
- Recruitment Intelligence: Ranks candidates with 92% match score using skills graph
- Attrition Prediction: Detects 47 high-risk employees 60 days before they leave
- Policy Assistant: Zero-hallucination HR Q&A using RAG on company policies
- Performance & OKR Intelligence: Connects Jira + reviews for 360-degree view
- Skill Graph: Maps 12,450 employees skills for internal mobility
- AI Interview Agent: Generates hard behavioral questions with real-time AI evaluation rubric (Technical Depth, Communication, Problem Solving, Culture Fit)
Main highlight: Our "Connected Telemetry Feeds Powering AI Reasoning" - We sync 6 systems (12k employee records, 18k attendance logs) and AI gives 3 layers: Multi-source Evidence, Business Impact ($380k cost calculation), and Prescribed HR Action.
How we built it
Frontend: React + TypeScript + Tailwind, deployed on Netlify at workmindai1.netlify.app AI Logic: Custom reasoning engine that correlates GitHub commits, attendance, surveys, performance data Data Layer: Simulated enterprise connectors for Workday, Greenhouse, BambooHR, Jira, Slack, Google Calendar Visualization: Recharts for headcount, attrition trends, department distribution AI Interviews: LLM-based evaluation with structured rubric - quantifying risk, ownership, RCA, mitigation
Team: Shama Parveen & Eishika Chauhan - We built all 6 modules in 48 hours with mobile-first design.
Challenges we ran into
- Connecting 6 different data schemas into one unified intelligence was tough - we made a normalization layer.
- Making AI evaluation objective, not subjective - so we added rubric key points like "quantifying risk objectively" and "ownership without blame"
- Showing $380k business impact calculation - we had to build a cost model for attrition.
- Netlify deployment with large dataset - optimized with lazy loading.
Accomplishments that we're proud of
- Built fully working platform with 6 live modules, not just a prototype
- Our Workforce Insights page shows real AI reasoning: Engineering Attrition Spike with evidence + impact + action - no other team did this
- AI Interview Agent that evaluates candidates on 4 dimensions in real-time
- 1,248 employees, 32 open positions, 480 engineering staff - all simulated but feels real
- Live link workmindai1.netlify.app is fully functional
What we learned
We learned that HR doesn't need more dashboards, they need prescriptive intelligence. Quantifying risk (like CVSS 8.5 for security or 17% attrition risk) makes HR talk to executives in business language. Also learned how to design AI that gives evidence, not just answers - to avoid hallucination.
What's next for WorkMind AI for Alexa+
- Integrate with Alexa+ for voice HR queries: "Alexa, who are my high attrition risks this month?"
- Real integration with Workday and Greenhouse APIs
- Add predictive hiring - which candidate will stay longest
- Slack bot for managers for daily workforce nudges
- Launch for startups with <500 employees as SaaS
Built With
- ai-reasoning-engine
- greenhouse-api
- llm
- netlify
- rag
- react
- recharts
- tailwind-css
- typescript
- workday-api