icetana AI launches Triage Agent to enhance surveillance alerts

Icetana AI has introduced the Triage Agent, an artificial intelligence-driven solution for improving the management of real-time surveillance for security teams.

The Triage Agent has been developed to help operators filter, prioritise, and receive the most important security alerts as they happen. By focusing on the most relevant incidents, the tool intends to reduce the risk of critical events going unnoticed amongst modern security systems' large volume of surveillance footage.

Kevin Brown, Chief Executive Officer of icetana AI, said, "Security operations teams are overwhelmed, often tasked with monitoring tens or even hundreds of camera feeds, with many alerts going unnoticed. Our Triage Agent allows a single operator to effectively oversee over 2,000 cameras, not just by increasing capacity but by ensuring precision, relevance, and response time."

The Triage Agent uses artificial intelligence to detect, validate, and categorise incidents according to criteria set by security professionals. This approach ensures that only significant events are passed on to operators in the control room, reducing manual review times and boosting operational efficiency.

The system integrates directly with icetana AI's broader product suite and bridges automated alerts and human judgment. Security teams can configure the Triage Agent to focus on specific incidents, such as weapon detection, property damage, or suspicious activities. The tool can automatically filter out footage deemed irrelevant while providing clear descriptions of each event that has been flagged for review by operators.

At the core of the Triage Agent is the use of large language models powered by GPT technology. These models assist in delivering context-aware and intelligent decision-making, which is intended to enhance the system's accuracy. When an anomaly is detected, the large language model processes the corresponding image based on the security team's predefined requirements. The model then examines the footage further to determine whether it should be escalated to human responders. This process of semantic filtering introduces a level of precision designed to improve upon conventional artificial intelligence workflows in surveillance security.

Brown added, "The Triage Agent functions as ten virtual security operators – constantly vigilant and never fatigued, ensuring that no critical event is missed. During testing, The Triage Agent identified all reportable incidents and even highlighted critical events that human operators had missed, proving its superior accuracy."

The introduction of Triage Agent is intended to support security operators as they manage large networks of surveillance cameras across locations such as public venues, transport infrastructures, and corporate facilities. The solution is presented as a means to reduce manual intervention while maintaining high levels of incident detection and response.

The Triage Agent can be tailored to the specific needs of each organisation, making it possible for operators to address unique risks and priorities by refining the system's attention on certain behaviours or criteria. By automating both the initial detection and the subsequent triage of events, icetana AI positions its new tool as a support to human teams, aiming to reduce workloads while attempting to protect against security lapses that can result from information overload.