	In Browser
	StumbleUpon
	del.icio.us
	Google
	Google Buzz
	reddit
	LinkedIn

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Attrition Risk Agent

Attrition Risk Agent

Author: Venkata Sudhakar

Losing a skilled employee costs ShopMax India 6 to 9 months of their salary in recruitment and productivity loss. Identifying at-risk employees before they resign allows the HR and management team to intervene with retention measures - a salary review, a promotion conversation, or a role change - before it is too late.

This tutorial builds a Gemini ADK agent that scores employees on attrition risk factors including tenure, recent performance trend, engagement score, salary competitiveness, and absenteeism, then flags high-risk employees with recommended retention actions.

The below example shows an employee attrition risk scoring agent for ShopMax India in a business context.

import asyncio
from google.adk.agents import LlmAgent
from google.adk.tools import FunctionTool
from google.adk.runners import Runner
from google.adk.sessions import InMemorySessionService
from google.genai import types

# Employee attrition signals
EMPLOYEES = [
    {"id": "EMP-501", "name": "Kavitha Menon", "role": "Software Engineer",
     "tenure_months": 14, "performance_trend": "declining",
     "engagement_score": 3.1, "salary_vs_market_pct": -12,
     "absent_days_last_quarter": 8},
    {"id": "EMP-502", "name": "Arun Prakash", "role": "Operations Manager",
     "tenure_months": 36, "performance_trend": "stable",
     "engagement_score": 4.2, "salary_vs_market_pct": 5,
     "absent_days_last_quarter": 2},
    {"id": "EMP-503", "name": "Pooja Desai", "role": "Marketing Executive",
     "tenure_months": 8, "performance_trend": "improving",
     "engagement_score": 2.8, "salary_vs_market_pct": -18,
     "absent_days_last_quarter": 5},
]

def score_attrition_risk() -> dict:
    results = []
    for emp in EMPLOYEES:
        risk = 0
        # Tenure: shorter tenure = higher risk
        if emp["tenure_months"] < 12:
            risk += 25
        elif emp["tenure_months"] < 24:
            risk += 15
        else:
            risk += 5
        # Performance trend
        if emp["performance_trend"] == "declining":
            risk += 20
        elif emp["performance_trend"] == "stable":
            risk += 5
        # Engagement score (out of 5, lower = higher risk)
        if emp["engagement_score"] < 3.0:
            risk += 25
        elif emp["engagement_score"] < 3.5:
            risk += 15
        else:
            risk += 5
        # Salary vs market
        if emp["salary_vs_market_pct"] < -15:
            risk += 20
        elif emp["salary_vs_market_pct"] < -5:
            risk += 10
        # Absenteeism
        if emp["absent_days_last_quarter"] > 6:
            risk += 10
        elif emp["absent_days_last_quarter"] > 3:
            risk += 5
        level = "HIGH" if risk >= 55 else ("MEDIUM" if risk >= 35 else "LOW")
        actions = []
        if emp["salary_vs_market_pct"] < -10:
            actions.append("Initiate salary review - below market rate")
        if emp["engagement_score"] < 3.5:
            actions.append("Schedule 1-on-1 stay interview with manager")
        if emp["performance_trend"] == "declining":
            actions.append("Discuss role challenges and provide coaching support")
        if emp["absent_days_last_quarter"] > 6:
            actions.append("Check wellbeing - high absenteeism may signal burnout")
        results.append({
            "employee_id": emp["id"],
            "name": emp["name"],
            "role": emp["role"],
            "risk_score": risk,
            "risk_level": level,
            "retention_actions": actions,
        })
    results.sort(key=lambda r: r["risk_score"], reverse=True)
    high_risk = sum(1 for r in results if r["risk_level"] == "HIGH")
    return {"total_assessed": len(results), "high_risk_count": high_risk, "employees": results}

attrition_tool = FunctionTool(func=score_attrition_risk)

agent = LlmAgent(
    name="attrition_risk_agent",
    model="gemini-2.0-flash",
    description="Scores ShopMax India employees on attrition risk and recommends retention actions.",
    instruction=(
        "You are an HR retention agent for ShopMax India. "
        "When asked for an attrition risk report, call score_attrition_risk. "
        "Present employees grouped by risk level: HIGH first, then MEDIUM, then LOW. "
        "For each HIGH risk employee, list specific retention actions and urgency."
    ),
    tools=[attrition_tool],
)

async def main():
    session_service = InMemorySessionService()
    runner = Runner(agent=agent, app_name="attrition_app", session_service=session_service)
    session = await session_service.create_session(app_name="attrition_app", user_id="hr1")
    msg = types.Content(role="user", parts=[types.Part(
        text="Run attrition risk assessment for all employees"
    )])
    async for event in runner.run_async(user_id="hr1", session_id=session.id, new_message=msg):
        if event.is_final_response() and event.content:
            print(event.content.parts[0].text)

asyncio.run(main())

It gives the following output,

Attrition Risk Report - ShopMax India

Assessed: 3 employees | HIGH Risk: 2

[HIGH RISK - 65] Pooja Desai | Marketing Executive
  Engagement: 2.8/5 | Salary: -18% vs market | Tenure: 8 months
  Actions:
    1. Initiate salary review - significantly below market rate
    2. Schedule 1-on-1 stay interview with manager this week

[HIGH RISK - 60] Kavitha Menon | Software Engineer
  Engagement: 3.1/5 | Salary: -12% vs market | Declining performance
  Actions:
    1. Initiate salary review - below market rate
    2. Schedule stay interview and discuss role challenges
    3. Check wellbeing - high absenteeism (8 days last quarter)

[LOW RISK - 20] Arun Prakash | Operations Manager
  Engagement: 4.2/5 | Salary: +5% vs market | Tenure: 36 months
  No immediate action required.

ShopMax India can run this attrition risk agent monthly and share the HIGH risk list confidentially with HR business partners and department heads. Early intervention - a salary correction or a candid career conversation - costs a fraction of the recruitment and training investment required to replace a departing employee.

Send your comments, suggestions or queries regarding this site to [email protected].