Balancing Security and Privacy: A Closer Look at Verkada’s Person of Interest Only Face Search Feature
As organizations adopt modern security technologies, they face a growing challenge: how to enhance safety while respecting privacy and meeting evolving regulatory expectations. Verkada addresses this need through a set of thoughtful privacy controls designed to support both security and compliance.
One such capability is Verkada’s Person of Interest Only Face Search (“POI Faces Only”) feature, which was built to help strengthen security outcomes while minimizing the collection and storage of personal data.
As the physical security market shifts toward more privacy-conscious design – particularly in regions like Europe and Australia where regulatory scrutiny around facial recognition technology continues to evolve – features like POI Faces Only are becoming increasingly relevant.
A targeted approach to facial recognition
A common security practice involves defining a list of specific individuals – known as persons of interest (POI) – such as banned individuals or known threats. This allows security teams to narrow their focus, prioritizing attention and alerts on individuals who may pose a higher risk.
In some cases, organizations may choose – or be required – to retain only POI-related data, while deleting biometric information associated with all other individuals. To help support this need, Verkada developed its POI Faces Only feature.
With POI Faces Only enabled, if a detected face does not match someone on the predefined POI list:
The facial data used for comparison is not stored
The image is not made searchable
No further processing occurs beyond the initial comparison
This means that only relevant, high-priority events are retained, significantly reducing unnecessary data collection. Additionally, POI Faces Only detection is processed directly on the camera, further enhancing privacy.
How Verkada’s POI Faces Only feature works
Verkada’s POI Faces Only feature is designed to perform key processing steps on the device itself, while maintaining a seamless user experience.
Here’s how it works in practice:
POI creation: An administrator creates a POI in Command based on an existing detection. A mathematical representation of that individual’s face (a “face vector”) is generated and securely stored in the cloud.
Device synchronization: When a camera comes online, it downloads the facial vectors to the device.
On-device detection: As individuals pass within view, the camera analyzes the facial vectors on the device itself.
Matching and alerts: The camera compares detected faces against the POI list it downloaded. If a match is found, an alert is triggered and relevant footage is flagged. The matched image is stored on the camera and sent to the cloud for analysis. Depending on the organization’s retention settings, it may be also backed up to the cloud.
Selective storage: Only footage associated with a confirmed POI match is stored and made searchable. All other detections are immediately discarded.
This architecture ensures that organizations are not sending every detected face to the cloud for analysis, nor are they retaining any footage of non-POI individuals.
Supporting privacy expectations
In markets where expectations around facial recognition technology continue to evolve, Verkada’s POI Faces Only feature offers a practical way to balance security needs with privacy considerations.
In Australia, for example, guidance from the Office of the Australian Information Commissioner (OAIC) emphasizes principles such as necessity, proportionality, and strong governance. At a high level, this means limiting the use of facial recognition to clearly defined purposes and avoiding the collection or retention of unnecessary personal data. Recent legal developments highlight that targeted use of facial recognition, paired with strict data minimization, can align with regulatory expectations.
Verkada’s POI Faces Only approach reflects this model. By restricting facial recognition to a predefined list of persons of interest and discarding all other detections, the feature helps organizations to focus on legitimate security needs while reducing broader privacy impact.
As regulations continue to evolve, particularly in regions like Europe and Australia, approaches that emphasize precision, restraint, and transparency will become increasingly important. Verkada’s POI Faces Only feature represents a meaningful step in that direction.
