Project Overview

Search behaviour has shifted: buyers now ask AI assistants before they buy. MentionHub answers the new question with numbers — not “where does my site rank” but “does AI mention my brand at all”.

The Challenge

AI answers are not deterministic. Asking once proves nothing, so the product needed a repeated sampling method that stands up when an agency presents it to a client.

Scope of Work

• Design of a statistical sampling method across multiple AI engines
• Analysis pipeline for mentions, ranking position, and sentiment
• Share-of-voice dashboard with competitor comparison
• White-label PDF report generator for agency use
• Multi-tenant architecture with a credit system and usage accounting

Key Features

• One prompt sampled repeatedly across several AI engines
• Detection of brand mentions, position order, and answer sentiment
• Share of voice against competitors in a single chart
• Full check history, so daily, weekly, and monthly trends are visible
• PDF reports carrying the agency logo and colours
• Operators plug in their own API keys, set their own rates, and sell capacity as credits

Technology

Laravel 12, Vue 3 with Inertia, and MySQL — built to be customised, scaled, and self-hosted.

Outcome

Agencies and marketing teams gain measurable proof of their optimisation work in the AI search era, and licensees can resell it as a subscription service.