AI Agent for Fleet Management: Vehicle Tracking, Maintenance Scheduling, and Driver Communication

The Fleet Management Challenge

Fleet operators with 50 to 5,000 vehicles face a constant operational balancing act. Every vehicle needs regular oil changes, tire rotations, brake inspections, and DOT compliance checks. Drivers need route updates, maintenance reminders, and emergency support. Managers need visibility into where every vehicle is, which ones are due for service, and which drivers are approaching hours-of-service limits.

An AI agent for fleet management ties together GPS telematics, maintenance rule engines, and communication channels into a single conversational interface that fleet managers and dispatchers can query naturally.

Modeling Fleet Vehicles and Maintenance Rules

Start with data models that capture vehicle state and maintenance requirements:

flowchart LR
    INPUT(["User intent"])
    PARSE["Parse plus<br/>classify"]
    PLAN["Plan and tool<br/>selection"]
    AGENT["Agent loop<br/>LLM plus tools"]
    GUARD{"Guardrails<br/>and policy"}
    EXEC["Execute and<br/>verify result"]
    OBS[("Trace and metrics")]
    OUT(["Outcome plus<br/>next action"])
    INPUT --> PARSE --> PLAN --> AGENT --> GUARD
    GUARD -->|Pass| EXEC --> OUT
    GUARD -->|Fail| AGENT
    AGENT --> OBS
    style AGENT fill:#4f46e5,stroke:#4338ca,color:#fff
    style GUARD fill:#f59e0b,stroke:#d97706,color:#1f2937
    style OBS fill:#ede9fe,stroke:#7c3aed,color:#1e1b4b
    style OUT fill:#059669,stroke:#047857,color:#fff

from dataclasses import dataclass, field
from datetime import datetime, date
from enum import Enum
from typing import Optional

class MaintenanceType(str, Enum):
    OIL_CHANGE = "oil_change"
    TIRE_ROTATION = "tire_rotation"
    BRAKE_INSPECTION = "brake_inspection"
    DOT_INSPECTION = "dot_inspection"
    TRANSMISSION_SERVICE = "transmission_service"

@dataclass
class MaintenanceRule:
    maintenance_type: MaintenanceType
    interval_miles: int
    interval_days: int
    description: str

@dataclass
class FleetVehicle:
    vehicle_id: str
    unit_number: str
    make: str
    model: str
    year: int
    current_mileage: int
    last_oil_change_miles: int
    last_oil_change_date: date
    latitude: float
    longitude: float
    speed_mph: float
    driver_name: str
    driver_phone: str
    status: str = "active"

MAINTENANCE_RULES = [
    MaintenanceRule(MaintenanceType.OIL_CHANGE, 7500, 180, "Engine oil and filter"),
    MaintenanceRule(MaintenanceType.TIRE_ROTATION, 10000, 365, "Rotate all tires"),
    MaintenanceRule(MaintenanceType.BRAKE_INSPECTION, 25000, 365, "Full brake check"),
    MaintenanceRule(MaintenanceType.DOT_INSPECTION, 0, 365, "Annual DOT compliance"),
]

GPS Tracking Integration Tool

The vehicle tracking tool simulates pulling real-time location data from a telematics provider like Samsara, Geotab, or Verizon Connect:

from agents import function_tool

FLEET_VEHICLES = [
    FleetVehicle("FV-001", "Unit 14", "Freightliner", "Cascadia", 2024,
                 142000, 135000, date(2025, 11, 15), 37.7749, -122.4194,
                 58.0, "Mike Torres", "+1-555-0101"),
    FleetVehicle("FV-002", "Unit 27", "Kenworth", "T680", 2023,
                 198000, 195500, date(2026, 1, 20), 34.0522, -118.2437,
                 0.0, "Sarah Kim", "+1-555-0102"),
    FleetVehicle("FV-003", "Unit 33", "Volvo", "VNL 860", 2025,
                 67000, 62000, date(2025, 12, 10), 41.8781, -87.6298,
                 62.5, "James Okafor", "+1-555-0103"),
]

@function_tool
def get_vehicle_location(unit_number: Optional[str] = None) -> str:
    """Get current GPS location and status for fleet vehicles."""
    vehicles = FLEET_VEHICLES
    if unit_number:
        vehicles = [v for v in vehicles
                    if v.unit_number.lower() == unit_number.lower()]

    if not vehicles:
        return "No matching vehicles found."

    lines = []
    for v in vehicles:
        status = "Moving" if v.speed_mph > 0 else "Stopped"
        lines.append(
            f"{v.unit_number} ({v.year} {v.make} {v.model}) | "
            f"Driver: {v.driver_name} | "
            f"Location: ({v.latitude:.4f}, {v.longitude:.4f}) | "
            f"Speed: {v.speed_mph} mph | Status: {status}"
        )
    return "\n".join(lines)

Maintenance Check Tool

This tool evaluates each vehicle against the maintenance rules and flags overdue or upcoming services:

@function_tool
def check_maintenance_status(unit_number: Optional[str] = None) -> str:
    """Check maintenance status for fleet vehicles based on mileage and time rules."""
    vehicles = FLEET_VEHICLES
    if unit_number:
        vehicles = [v for v in vehicles
                    if v.unit_number.lower() == unit_number.lower()]

    today = date.today()
    alerts = []

    for v in vehicles:
        for rule in MAINTENANCE_RULES:
            miles_since = v.current_mileage - v.last_oil_change_miles
            days_since = (today - v.last_oil_change_date).days

            overdue_miles = (rule.interval_miles > 0
                            and miles_since >= rule.interval_miles)
            overdue_days = days_since >= rule.interval_days

            if overdue_miles or overdue_days:
                reason = []
                if overdue_miles:
                    reason.append(f"{miles_since} miles since last service")
                if overdue_days:
                    reason.append(f"{days_since} days since last service")
                alerts.append(
                    f"OVERDUE: {v.unit_number} needs {rule.description} "
                    f"({', '.join(reason)})"
                )

    return "\n".join(alerts) if alerts else "All vehicles are current on maintenance."

Driver Notification Tool

@function_tool
def send_driver_message(
    unit_number: str,
    message: str,
    priority: str = "normal",
) -> str:
    """Send a message to a fleet driver via their registered phone number."""
    vehicle = next(
        (v for v in FLEET_VEHICLES
         if v.unit_number.lower() == unit_number.lower()), None
    )
    if not vehicle:
        return f"Vehicle {unit_number} not found in fleet."

    # In production, call Twilio / SMS API here
    return (
        f"Message sent to {vehicle.driver_name} ({vehicle.driver_phone}): "
        f"[{priority.upper()}] {message}"
    )

Assembling the Fleet Agent

from agents import Agent, Runner

fleet_agent = Agent(
    name="Fleet Manager",
    instructions="""You are an AI fleet management assistant. You can:
    1. Track vehicle locations and speeds in real time
    2. Check maintenance schedules and flag overdue services
    3. Send messages to drivers with normal or urgent priority
    Always prioritize safety-related maintenance alerts.""",
    tools=[get_vehicle_location, check_maintenance_status, send_driver_message],
)

result = Runner.run_sync(
    fleet_agent,
    "Which vehicles have overdue maintenance? Notify those drivers."
)
print(result.final_output)

FAQ

How do I integrate with real GPS telematics providers?

Most providers like Samsara, Geotab, and KeepTruckin offer REST APIs. Replace the in-memory fleet list with API calls that fetch live vehicle positions. Use webhook subscriptions for real-time event streaming instead of polling, and cache location data for 30 to 60 seconds to reduce API costs.

Can the agent handle hours-of-service (HOS) compliance?

Yes. Add a tool that queries ELD (Electronic Logging Device) data for each driver. The tool checks remaining drive time, mandatory break requirements, and 70-hour weekly limits. If a driver approaches a threshold, the agent can proactively alert dispatch to plan a relief driver or rest stop.

How should I handle maintenance scheduling conflicts?

In production, integrate with a shop management system that tracks bay availability and technician schedules. The agent should check open slots before scheduling and offer the driver the nearest available time. Use optimistic locking on appointment slots to prevent double-booking.

---

#FleetManagement #VehicleTracking #MaintenanceAI #Logistics #Python #AgenticAI #LearnAI #AIEngineering