Prarambh AI

Inspiration

Employee onboarding is often broken. HR teams are overwhelmed with paperwork, new hires feel lost and unproductive for weeks, and mentors struggle to find time to guide them effectively. We realized that the traditional "one-size-fits-all" approach to onboarding simply doesn't work. We were inspired to build Prarambh AI to fix this by creating a personalized, intelligent, and seamless onboarding experience that gets every new hire ready from day one, while significantly reducing the operational burden on HR.

What it does

Prarambh AI is an intelligent Onboarding OS that manages the entire lifecycle of a new hire's integration into a company. Key features include:

• AI Task Builder: Generates role-specific onboarding plans in seconds by simply analyzing a job description.
• Smart Buddy Matching: Automatically pairs new hires with the most appropriate mentor based on role, skills, and experience.
• Document Intelligence: Upload HR policies and handbooks, and our AI automatically extracts relevant onboarding tasks and milestones.
• Live Analytics & Compliance Tracking: Provides real-time dashboards for HR to track completion rates, engagement scores, and audit trails for compliance.
• 24/7 AI Assistant: A trained chatbot that understands company policies, answers new hire questions instantly, and escalates complex issues when necessary.
• Integration Hub: Automatically provisions access to essential tools like Slack, GitHub, Notion, and more on day one.

The platform provides dedicated, tailored experiences for four key roles: Admins, HR Managers, Mentors, and New Hires.

How we built it

We architected Prarambh AI to be fast, responsive, and highly intelligent:

• Frontend: We built a highly interactive and beautiful user interface using React 18, TypeScript, and Vite. For styling, we used Tailwind CSS combined with Framer Motion for smooth, dynamic animations and a premium glassmorphism aesthetic.
• Backend & AI Engine: The core intelligence is powered by Python (FastAPI), LangGraph, and MongoDB. We integrated the Gemini API as our foundational language model to power the AI Task Builder, Document Intelligence, and the 24/7 AI Assistant.
• Design System: We crafted a dark-mode optimized design with a custom DotField component for interactive visual effects, ensuring the platform feels modern and engaging.

Challenges we ran into

One of the main challenges was parsing unstructured HR documents and reliably extracting actionable onboarding tasks without hallucinating. We had to carefully engineer the prompts and utilize LangGraph to create a robust, multi-step reasoning pipeline for the AI. Additionally, creating a unified platform that felt intuitive for four distinct user personas (Admin, HR, Mentor, New Hire) required significant UX planning and state management on the frontend.

Accomplishments that we're proud of

We are incredibly proud of the AI Task Builder. Seeing the platform take a simple job description and instantly generate a comprehensive, week-by-week onboarding plan is magical. We are also proud of the premium look and feel of the application; the micro-animations and glassmorphic UI make enterprise software feel like a consumer app.

What we learned

Building Prarambh taught us a lot about the nuances of agentic AI workflows. We learned how to chain LLM calls effectively using LangGraph to achieve more complex and reliable outcomes than a single prompt could provide. We also deepened our knowledge of framer-motion to create fluid, accessible animations in React.

What's next for Prarambh AI

Moving forward, we want to expand our Integration Hub to support automated provisioning for over 50+ SaaS applications. We also plan to introduce predictive analytics to help HR identify employees who are at risk of churning during their first 90 days, allowing for proactive intervention. Finally, we want to add multi-language support to the AI Assistant for global remote teams.

Built With

• fastapi
• framer-motion
• langgraph
• mongodb
• python
• react
• tailwind-css
• typescript
• vite