XEdgAI - IntentSentinel
Surveillance cameras are everywhere, but most systems still serve mainly as tools for post-incident evidence and investigation. They can show what happened, but rarely help prevent an incident while it is unfolding.
IntentSentinel changes this by combining object detection, multi-object tracking, and vision-language understanding on a single industrial edge AI system. It goes beyond alerts such as “a person entered the area” to identify behaviors like someone trying multiple car door handles, placing merchandise into a personal bag, or tailgating through an access-controlled door.
Potential threats are immediately sent, with visual evidence and reasoning, to security staff for human confirmation and response. All actions are recorded in a tamper-evident audit trail.
Everything runs locally on-site, with no cloud dependency and no biometric database.
From post-incident investigation to real-time prevention.