Production mobile app on both app stores. AI replaces manual vehicle inspections for fleet operators.
Large fleets process 50 to 200 vehicle inspections daily. Manual reports take hours and create bottlenecks at shift changes. The same scratch gets rated differently by different inspectors depending on their experience, their mood, and whether they are rushing to finish before lunch. Reports pile up, managers fall behind, and damage that should have been caught early becomes expensive repair jobs weeks later. The inspection process was the bottleneck for every fleet operation downstream.
Driver takes photos of the vehicle. The AI classifier analyzes each image, detects damage types including scratches, dents, broken windows, and paint deterioration, assigns severity levels, and generates a structured report automatically. No manual review needed before reports reach fleet managers. The same scratch gets the same severity rating regardless of which driver photographs it. Consistency that human inspectors cannot provide at scale.
Each fleet operator gets their own isolated environment. Data isolation enforced at the API level with tenant-scoped tokens. Drivers see only their assigned vehicles. Fleet managers see the full dashboard with current vehicle status, most recent inspection, and damage history over time. Admins control access and onboarding. Adding a new company requires a configuration entry, not schema changes. Clean Architecture in Flutter separates presentation, domain, and data layers. The .NET Core backend handles authentication, fleet organization, photo storage, and the AI classification pipeline.
Single Flutter codebase deploys to iOS and Android with native access to camera, file system, and push notifications. Live on App Store and Google Play since August 2024, handling real fleet operations with real users. PostgreSQL stores the complete vehicle lifecycle: inspection records, damage assessments, user permissions, and fleet configurations, all tenant-scoped at the database level.
Have something like this?
Let's talk