The Detection Reviews feature of the Blindsight portal allows you to label the risk level of each detection and report AI Errors. This has several benefits and is discussed further in this help article. However, depending on the volume of detections produced by your Blindsight devices and your organisation's priorities, it may not be practical to manually review every detection.

Presien's AI Risk Assessment is intended to help you get a profile of risk across your organisation without the need for large scale manual reviews. With AI Risk Assessment enabled every critical, person detection sent to the cloud by your Blindsight devices is automatically assessed and assigned a risk level based on a combination of observable risk factors including:

This automatic assessment is imperfect and should not be relied upon in place of expert opinion of trained personnel, but aims to achieve some of the benefits of risk based detection reviews without the need for large scale manual review.

AI Risk Assessment results will be visible on the detection card when available, displayed as an AI Suggestion on the corresponding risk button. You can choose to agree or disagree with the AI Suggestion by selecting the risk level you wish to apply.

For the sake of reporting and filtering the risk value associated with the detection will be the most recent review. Prior to manual review AI Suggested risk levels will be used to categorise the event. After manual review the most recent manual review will be used.

Assessing risk is complex and heavily dependant on operational context. Our goal is to help you understand risk across your operations in terms that are appropriate to your context. When you manually review the risk level of your detections your feedback is used to improve AI Risk Assessment and better calibrate its outputs to your expectations. If you have any feedback on this feature or would like to learn more about how it works please email us at support@presien.com.