Home/Case Study
Case Study

HR Agent for a 4,000-Employee Retailer

An AI-powered assistant that handles employee questions, conducts exit interviews, and manages surveys via WhatsApp.

The Challenge

The HR team at a major Turkish retailer was overwhelmed. With 4,000 employees spread across multiple locations, they faced constant questions about salaries, leave policies, labor laws, and internal procedures.

Traditional approaches had clear limitations:

  • No support outside business hours
  • Inconsistent answers depending on who you asked
  • Exit interviews were superficial, missing the real reasons for turnover
  • Manual survey distribution was time-consuming and hard to track

Our Solution

We built a WhatsApp-based AI assistant that employees could message anytime. The system uses a modular architecture with specialized capabilities:

Conversational AI

LLM-powered chat with intent detection for 7 different query types: salary, leave, procedures, labor law, and more.

RAG Knowledge Base

100+ pages of HR documents and Turkish labor law indexed with vector embeddings for accurate retrieval.

Dynamic Exit Interviews

Adaptive questioning that digs deeper based on responses, uncovering root causes traditional forms miss.

Survey Automation

Bulk distribution, progress tracking, and Excel export for company-wide pulse surveys.

Tech Stack

GPT-4 / Claude 3.5 SonnetRAG with FAISS vector storeWhatsApp via Twilio APINext.js Admin DashboardPostgreSQL + Flask Backend

Results

  • 24/7 availability for employee questions
  • Scalable support without growing HR headcount
  • Consistent, up-to-date information across the organization
  • Deeper exit interview insights through adaptive questioning
  • Automated survey distribution and tracking

"Turkey's first WhatsApp-based LLM+RAG HR assistant, providing 24/7 employee support, dynamic exit interviews, and automated surveys."

Want something similar for your team?

Let's discuss how we can build an AI solution tailored to your needs.

Get in Touch