Functional DemoSaaS

Driver Retention Risk Engine

Predictive workforce-intelligence SaaS that flags which truck drivers are at risk of quitting, why, and what to do.

Retention Engine dashboard for Summit Freight Lines: KPI tiles for monitored drivers, critical and high risk counts, a drivers-needing-action list with risk scores, and top risk categories. — Driver retention dashboard

What it solves

Driver turnover in trucking is expensive, disruptive, and often preventable, but drivers rarely quit over one issue — they leave after repeated friction across pay, home time, dispatch, and workload. Fleets lack an early-warning layer that connects these signals and recommends action before a driver leaves.

Built for Trucking and logistics fleets — fleet owners, operations managers, retention and dispatch managers, HR, and safety/compliance teams.

Overview

Driver Retention Risk Engine is an intelligence layer that sits above a fleet's existing TMS, ELD, telematics, payroll, CRM, and HR systems to predict which drivers are at risk of quitting, explain why, and recommend practical retention actions. It scores risk from patterns like inconsistent pay, poor home time, dispatch friction, detention, idle time, and workload. Fleet and HR managers use it to intervene before turnover happens. Nightly jobs recompute risk and surface drivers who need attention.

What this demonstrates

A Next.js application with role-based access and encrypted integration credentials that computes driver risk scores and explanations over a Prisma/Postgres model. It runs scheduled and queued background work (a worker plus a nightly cron), imports data via CSV, and is covered by unit and Playwright end-to-end tests.

What makes it different

Know which drivers are about to quit, why they are at risk, and exactly what to do before they leave — an intelligence layer that sits above your TMS, ELD, payroll, and HR systems, not a replacement for them.

Intelligence layer, not a replacementSits above TMS, ELD, telematics, payroll, CRM, maintenance, HR, and messaging systems.
Friction patterns, not single issuesDrivers quit from repeated patterns — pay, home time, dispatch, detention — not one isolated event.
Explainable, not black-boxEvery signal includes evidence and a plain-English explanation; no unexplained flags.
Combine sentiment with operational evidenceThe system never overreacts to one phrase; sentiment is corroborated by operational data.
Calm and operationally usefulSimple and actionable, with no analytics overload.
Detection into actionTurns every risk into a practical, assignable retention intervention.

Core capabilities · 12

Integration Layer
Universal adapter framework for API, CSV, SFTP, webhook, and scheduled syncs across eight system categories.
Data Normalization Layer
Normalizes and deduplicates scattered source data, mapping external IDs across systems.
Unified Driver Timeline
Single timestamped, source-labeled event timeline queryable by driver, date, category, and severity.
Risk Signal Detection
Generates nine signal types with severity, confidence, evidence, and explanation.
Driver Risk Scoring Engine
Weighted, explainable score with category breakdown, trend, and preserved history.
Intervention Recommendation Engine
Recommends and tracks retention actions with priority, outcome, and risk before/after.
Facility Friction Intelligence
Groups stops by facility to surface detention, complaints, and churn-correlated locations.
Dispatcher & Manager Health View
Aggregates driver risk by dispatcher and manager, framed as coaching not blame.
Admin Rules & Configuration
Tunable risk weights and thresholds with versioned, validated changes.
Notifications & Alerts
Deduplicated, prioritized alerts for critical risk, spikes, overdue interventions, and integration failures.
Reports & Executive Insights
Fleet, at-risk-driver, facility, dispatcher, and intervention-performance reports with CSV/PDF export.
Intelligence Improvement Loop
Learns from resignations and intervention outcomes to improve future recommendations.

How the risk score works

Driver Risk Score (0–100: Low / Moderate / High / Critical)

Pay Risk20%
Pay decline, variance, unpaid detention, and disputes.
Home-Time Risk18%
Missed home time and long stretches away.
Dispatch Friction15%
Route changes, rejections, and dispatcher conflict.
Workload Imbalance12%
Over/under-utilization and unfair load assignment.
Detention Risk10%
Long shipper/receiver waits and repeat bad facilities.
Fatigue / HOS Pressure10%
High drive density, near-violations, and poor rest.
Equipment Risk7%
Truck downtime, repeated defects, and unresolved issues.
Sentiment Risk5%
Frustration, disengagement, and resignation language.
Safety / Compliance Pressure3%
Safety and compliance pressure indicators.

How it works

Connect source systems: TMS, ELD, payroll, CRM, messaging, maintenance, HR
Normalize the data into a unified driver timeline
Detect risk signals with evidence across pay, home time, dispatch, and more
Score each driver — weighted and explainable — and rank by risk
Surface the highest-risk drivers, why, and what to do today
Recommend and assign interventions
Track intervention outcomes and risk before/after
Learn from resignations vs. retained to improve the model

Key screens

Dashboard
Driver Risk List
Driver Risk Profile
Risk Timeline
Intervention Center
Facility Friction Map
Dispatcher Health View
Integration Settings
Admin Rules
Reports

Data model

Organization
User
Driver
Load
DriverActivityEvent
RiskSignal
DriverRiskScore
Intervention
Facility
IntegrationConnection

Screens · 6

Interventions table listing 'Review this driver's pay' actions per driver with priority, status (recommended, open, completed), assignee, and due date columns. — Interventions queue

Integrations page with connected sample data sources (TMS, ELD/HOS, Payroll, CRM) and a grid of coming-soon native integrations like Samsara, Motive, and Geotab. — Integrations

Driver risk-detail view with a recommended pay-review action, a weighted risk breakdown bar list (pay, home time, dispatch friction), and active risk signals with dated evidence. — Driver risk detail with evidence

Retention Engine dashboard showing 24 monitored drivers, 2 critical and 5 high risk, 6 open interventions, a drivers-needing-action list, and a top risk categories panel. — Retention dashboard overview

Drivers list for 24 monitored drivers with columns for terminal, dispatcher, tenure, trend, and a color-coded risk score badge per driver. — Monitored drivers list

How it is built

A single Next.js app (App Router) with server-side jobs, Auth.js authentication, and a Prisma + Postgres database via Docker Compose; background processing runs inline in dev and queued in production, optionally Redis-backed.

APIs and services

Auth.js

Classification

Industries: Logistics
Capabilities: Analytics Authentication Reporting Dashboards Background Job Pipelines Multi-Tenant Architecture Workflow Automation
Technologies: Docker Next.js Node.js PostgreSQL React Tailwind CSS TypeScript Prisma Zod

Building something with similar Analytics, Authentication, Reporting functionality?

Start a project and reference this one — the conversation begins with what you have already seen.

Start a project