Project Overview
Fleetwise AI is an advanced fleet diagnostic and repair management platform designed for heavy-duty truck service shops. It combines AI-powered diagnostics, automated work order processing, and knowledge management into a single unified system. By integrating conversational AI, intelligent data extraction, and real-time analytics, the platform significantly reduces diagnostic time, improves repair accuracy, and unlocks new revenue streams through automated warranty recovery. Built with a scalable multi-tenant architecture, Fleetwise AI serves as a centralized operating system for modern fleet service businesses.
Overview
Key Challenges
Heavy-duty truck service shops faced major inefficiencies due to manual diagnostics, fragmented systems, and lack of centralized data. Technicians spent 45–60 minutes diagnosing each vehicle, tribal knowledge was lost with senior staff turnover, and shops missed $30K–$80K annually in warranty recovery due to manual processes. Disconnected tools created workflow friction and operational inefficiencies.
Our Solution
We developed Fleetwise AI, an end-to-end SaaS platform integrating conversational AI diagnostics, automated work order processing, and a centralized knowledge base. The system uses GPT-powered assistants, AI-based PDF extraction, and intelligent warranty analysis to streamline operations, reduce manual effort, and provide real-time insights across multi-tenant environments.
Project Performance Metrics
Technical Details
Technical Architecture
Technical Features
Security Measures
- Row-Level Security (RLS) for tenant isolation
- JWT-based authentication with role-based access control
- Security-definer functions to prevent RLS recursion
- Secure AI processing with retry and timeout handling
Scalability Features
- Multi-tenant architecture with company-level isolation
- Hot-swappable AI model configuration
- Serverless Edge Functions for scalability
- Retry logic with exponential backoff for AI operations
Testing Coverage
Detailed Solution
An AI-first, multi-tenant SaaS platform designed to automate diagnostics, capture technician knowledge, and unlock hidden revenue through intelligent workflows.
AI-Powered Diagnostic Operating System
- •Conversational AI assistant guiding technicians step-by-step through diagnostics
- •Context-aware intelligence using truck history, fault codes, and repair data
- •Automatic capture of readings, parts, and actions from natural conversations
- •Voice-enabled workflows using speech-to-text for hands-free operation
- •Centralized system replacing fragmented tools and manual processes
Multi-Tenant SaaS Infrastructure
- •Company-level data isolation using Row-Level Security (RLS)
- •Supports multiple shops, teams, and locations
- •Scalable serverless architecture using Supabase
- •Secure authentication and role-based access control
Performance & AI Reliability
- •AI retry mechanisms with timeout handling
- •Progressive response streaming for better UX
- •Edge Functions for secure and scalable processing
- •99.7% uptime with production-grade stability
Work Order Automation Engine
- •AI-powered PDF extraction converting work orders into structured data
- •Automatic extraction of VIN, fault codes, and customer information
- •Auto-creation of truck profiles using VIN decoding
- •Eliminates manual data entry and reduces human errors
Warranty Intelligence System
- •AI engine detecting missed warranty opportunities
- •Cross-referencing manufacturer coverage and repair timelines
- •Identifies recoverable revenue automatically
- •Transforms hidden losses into measurable profit
Knowledge Capture & Learning System
- •Centralized knowledge base for technician-submitted fixes
- •Two-level approval workflow for quality assurance
- •Injected into AI context for smarter diagnostics
- •Preserves tribal knowledge across teams
AI Orchestration Layer
- •Multi-model AI system (OpenAI, Claude, lightweight models)
- •Dynamic model selection based on task complexity
- •Retry logic and fallback mechanisms for reliability
Real-Time Workflow Engine
- •End-to-end workflow automation from diagnosis to repair completion
- •Task-based system for technicians and managers
- •Centralized tracking of vehicle lifecycle
Implementation & Infrastructure
AI-driven, multi-tenant SaaS architecture built with serverless infrastructure, real-time systems, and multi-model AI orchestration.
AI-Driven SaaS Architecture (SPA + Serverless)
Fleetwise AI is built as a client-side React SPA powered by Supabase backend services, including PostgreSQL, authentication, and Edge Functions. The architecture is optimized for multi-tenant SaaS, enabling secure company-level data isolation while supporting AI-driven workflows and real-time operations at scale.
Core Technical Decisions
- Multi-tenant architecture using Row-Level Security (RLS) for strict data isolation
- AI model orchestration with support for OpenAI, Claude, and lightweight models
- Retry logic with exponential backoff for reliable AI responses
- Progressive loading states to improve AI interaction experience
- Secure Edge Functions for handling sensitive operations like OCR and user management
- JSONB-based flexible schema for handling complex truck specifications
Layered System Architecture
The platform follows a layered architecture designed for scalability, AI integration, and real-time performance across multiple service locations.
Presentation Layer
- React 19 SPA with role-based routing
- Radix UI + Tailwind for accessible UI
- Mobile-first technician experience
- Voice-enabled interaction support
API & AI Layer
- Supabase Edge Functions (Deno)
- AI diagnostic engine (GPT / Claude)
- OCR pipeline for work order extraction
- Warranty analysis engine
- External APIs for fault codes & research
Data Layer
- PostgreSQL with 20+ structured tables
- JSONB fields for flexible technical data
- Row-Level Security (RLS) policies
- Indexed queries for performance
- Company-level data isolation
AI Performance & Reliability
- Multi-model AI execution based on task complexity
- Retry mechanisms with timeout and failure handling
- Progressive feedback during long AI operations
Security & Access Control
- JWT-based authentication with Supabase Auth
- Role-based access across 5 distinct user roles
- SECURITY DEFINER functions to prevent RLS recursion
Gallery
Visual transformation showcasing the dramatic improvements across all interfaces
"Fleetwise AI has completely transformed how our shop operates. Diagnostics are faster, data is centralized, and we’re recovering revenue we didn’t even know we were losing."
Fleet Service Operator
Shop Owner
Project Details
Duration
8 months (Ongoing Development)
Industry
Fleet Management & Commercial Vehicle Repair
Team
Project Lead / Full-Stack Architect, Frontend Developer, Backend Developer, AI/ML Engineer, UI/UX Designer, QA Engineer
Technologies Used
Project Goals
- Reduce diagnostic time using AI-guided workflows
- Eliminate manual work order entry through automation
- Capture and preserve tribal knowledge
- Recover missed warranty revenue opportunities
- Enable multi-tenant SaaS for multiple shop locations
- Ensure high system reliability with AI fallback mechanisms
