How it works
The HR system (Workday, HubSpot, or Zapier) changes the employee's status — the bot instantly messages them on Telegram, Slack, or WhatsApp. It explains why the interview matters and guarantees anonymity: only HR sees the data, not the direct manager. The employee picks a reason for leaving, records a voice message or types. The AI asks 2–3 follow-up questions based on the responses and asks: can we retain you? It generates a summary and sends it to HR — only then does offboarding begin.
The bot explains why the interview matters and guarantees anonymity upfront — removing the barrier to honest feedback. The employee picks from 6 departure reasons, the AI probes deeper, and finally asks: if we fixed this, could we keep you? That retention signal reaches HR before the last working day.
The AI runs the interview adaptively — each question is shaped by the previous answer. Voice messages are transcribed, key words become the basis for follow-up questions. This is not a form — it's a live dialogue that surfaces real reasons.
What is assessed:
❓ Primary reason for leaving (6 categories)
✅ AI probes: was it a specific trigger, long-building tension, or a sudden incident?
❓ Voice or text explanation — in their own words
✅ AI transcribes, analyses tone, and asks 2–3 targeted follow-ups
❓ Could the employee be retained if the issue were resolved?
✅ Retention signal captured as a separate field — HR sees it immediately
❓ What they valued about the company; what they'd recommend improving
✅ AI produces a brief summary and risk score for HR
HR sees all exit interviews in one table: reason, retention signal, AI summary, and completion status. Patterns emerge across the quarter without manual analysis — recurring reasons highlight systemic issues.
Talpha