Ad-MOTO — Mobile Video Advertising Platform

The Brief

Ad-MOTO is an AdTech company that puts digital screens on electric motorbikes and delivers video advertising to urban audiences as the bikes move through city streets. They came to us needing the full technology stack: fleet management, device communication, content delivery, audience measurement, campaign analytics, and rider operations — all working together in real time over cellular connections.

This wasn’t a simple web app. It was a platform that had to communicate with hardware on moving vehicles, push video content to Raspberry Pi screens over variable 4G connections, and provide advertisers with quantifiable audience metrics.

The Approach

We built the entire platform as a suite of interconnected Laravel applications. MQTT was chosen for device communication — lightweight, reliable over patchy cellular, and well-suited to the publish/subscribe pattern needed for fleet-wide content pushes. Each bike’s Raspberry Pi subscribes to MQTT topics for content updates and publishes telemetry (screen status, connectivity, playback confirmation) back to the platform.

The architecture splits into two main systems: the Hub (fleet operations, device management, campaign scheduling) and the Portal (rider onboarding, document management, shift tracking). Both share a common data layer but serve very different user groups with different needs.

The Build

The Hub handles programmatic ad scheduling with slot-based conflict resolution, real-time fleet tracking via Flespi GPS, and live dashboards showing every bike’s location, screen status, and current content — all updating via WebSockets without page refresh. Video content uploaded by advertisers is transcoded via PHP-FFmpeg to formats optimised for the screen hardware and compressed for cellular delivery.

Audience measurement uses Wi-Fi probe detection from the screen units to estimate unique devices near each bike, feeding heatmap visualisations that show advertisers exactly where their campaigns reached and how many people were nearby.

The Portal manages rider recruitment with AI-powered document verification — riders upload their documents and Claude AI extracts key fields, cross-references against profile data, and assigns confidence scores. High-confidence submissions are auto-approved; edge cases go to a human review queue. This replaced a manual process that was becoming a bottleneck as the fleet scaled.

We also built a natural language reporting feature: campaign managers ask questions in plain English and receive generated reports with visualisations. No SQL knowledge required to interrogate complex advertising performance data.

The Result

Ad-MOTO runs their entire operation — fleet management, content delivery, rider onboarding, campaign analytics, and client reporting — from the platform we built. The MQTT architecture reliably handles communication with devices on moving vehicles. The AI document verification reduced rider onboarding time significantly. Campaign managers can self-serve their own reporting without developer involvement.