AI Attrition Prediction: No More Resignation Surprises
The cost of losing one key employee is estimated at 50–200 % of their annual salary (recruitment, onboarding, knowledge loss, temporary productivity drop). For a company with 30 employees, 3 departures a year represent a real cost of 50,000–100,000.
Signals AI Reads
Research shows that an employee decides to leave 60–90 days before submitting their notice. During this period, behaviour is measurably different:
Attendance:
- Increase in lateness and short-term absences
- Large block of annual leave taken at once
- Higher number of sick days
Performance:
- Decline in goal or KPI achievement
- Less initiative in projects
- Shorter average working day
Engagement:
- Low scores in internal pulse surveys
- Reduced interaction on team platforms
- Unused training or certifications
Additional contextual factors:
- Employee was passed over in the last promotion cycle
- Change of manager
- Declining morale in the team
Output: Monthly HR Risk Report
Employee Risk Profile — April 2026
🔴 High risk:
- Jana K., Senior Sales — score 82/100
Signals: +40 % absences, 3 unused benefits,
no Christmas bonus this year
🟡 Medium risk:
- Marek V., IT Specialist — score 58/100
Signals: declining performance, 15 days holiday taken at once
✅ Stable (score <40): 24 employeesRecommended Actions
For each at-risk employee the system suggests:
- Timing of the conversation (when is most appropriate)
- Topics to cover (what has changed, what is missing)
- Available retention levers (salary increase, remote work, project, mentoring)
ROI: One Retained Employee = 20,000–80,000 Saved
For an IT specialist with a salary of 50,000 €/year:
- Cost of departure: ~75,000 € (1.5× salary)
- Cost of retention conversation + potential 10 % raise: 5,000 €/year
- Net saving of one retained employee: 70,000 €
