The technology that powers the Intelligent Workspace.

Frequently Asked Questions

• What is the CogniPoint sensor and how does it work?

  CogniPoint is PointGrab’s optical edge-AI sensor — an intelligent device that combines an image sensor with on-chip AI processing to detect and analyze workspace occupancy. The AI model runs entirely on the sensor itself, which means all processing happens at the edge, inside the device. No images or video are ever transmitted or stored. CogniPoint can accurately count people, track movement patterns, detect objects, and monitor traffic flow — all in real time and from a single device.

• How is CogniPoint different from PIR or thermal sensors?

  There are three main occupancy sensor technologies on the market, and they differ significantly in capability:

  PIR (Passive Infrared) detects motion via changes in infrared radiation. It is inexpensive and widely deployed, but it can only detect motion — it cannot count people or detect stationary occupants, which makes it unreliable for real occupancy intelligence.

  Thermal sensors detect body heat signatures and are often marketed on a privacy-first narrative since they capture no visual data. However, thermal technology is physically limited in resolution, performs poorly in warm environments, and struggles with desk-level accuracy.

  Optical Edge-AI (CogniPoint) combines camera-based scene capture with on-chip AI processing. This delivers precise headcounts, sub-meter positioning, desk-level granularity, and object detection — with full privacy protection, because no images ever leave the device. CogniPoint resolves the core limitations of both PIR and thermal while delivering a level of accuracy and detail that neither can match.

• Does CogniPoint capture, transmit, or store images?

  No. CogniPoint is a privacy-by-design sensor. While it uses an optical sensor to perceive the environment, all AI inference runs entirely on the chip inside the device. The sensor never transmits, stores, or exposes images or video of any kind. What leaves the sensor is purely anonymized occupancy data — counts, positions, and movement indicators. This makes CogniPoint suitable for privacy-sensitive environments, including offices in GDPR-regulated regions.

• What types of occupancy data does CogniPoint provide?

  CogniPoint delivers four core data types from a single device:

  Occupancy Count — accurately detects whether seats and spaces are occupied, and how many are actually there.

  People in Motion — monitors the anonymous location of people within a defined detection area.

  Object Detection — identifies physical objects such as desks, chairs, and other furniture to verify that your floor plan matches reality. Also, can detect objects such as laptops, smartphones, bags coats, etc. for passive occupancy in spaces.

  Virtual Traffic Line — tracks the number of people entering and exiting a space, providing total people count for larger areas such as lobbies or open floors, or for restrooms for demand-based cleaning applications.

• How accurate is the data, and how frequently is it updated?

  CogniPoint achieves 95%+ accuracy in occupancy detection, with data processed 600 times a second and transmitted at one-minute intervals and a latency of under one second. This combination of high accuracy and near real-time updates makes the data reliable enough to drive automated decisions — from space booking systems and HVAC controls to cleaning schedules and capacity planning.

• What is the difference between the wired and wireless CogniPoint sensors?

  PointGrab offers both wired and wireless versions of CogniPoint to accommodate different workspace environments and installation requirements. Wired sensors are ideal for new buildings, refurbishment projects and more permanent installations with structured cabling infrastructure, while wireless sensors offer greater flexibility for existing buildings and short term surveys where cabling is not practical. Both versions share the same edge-AI core and deliver the same data quality and feature set.

• What is the PointGrab Cloud Management Platform?

  The PointGrab Cloud Management Platform is the software layer that sits above the sensors and turns raw occupancy data into actionable intelligence. Through a flexible UI, you can define custom detection areas and counting zones independently of sensor placement. The platform supports stitched areas — combining data from multiple sensors to deliver wider-area coverage — as well as smaller Areas of Interest, which allow you to drill down into highly localized insights at the desk or zone level. All data is aggregated, cleaned, and made available for export or direct integration through a standard RESTful-API.

• Can PointGrab integrate with our existing workplace applications?

  Yes. PointGrab is built for open integration. The platform exposes a standard REST API, and the occupancy data it outputs is software-agnostic — meaning it can feed directly into any workplace management system, building management platform, analytics dashboard, booking system or custom application you already use. Whether you need to connect to a desk booking tool, an HVAC controller, a utilization analytics platform, or a facilities management system, the integration path is straightforward.

• How many sensors do I need to cover my space?

  The number of sensors may depend on the size of the space, furniture layout, ceiling heights and the granularity of data you need. CogniPoint is designed for flexible deployment — a single sensor can cover a defined zone, and the Areas of Interest feature splits the data into the relevant zones seamlessly so there are no coverage gaps. Our team can help you determine the right sensor density for your specific floor plan and use case during the onboarding process.

• What kinds of spaces is CogniPoint designed for?

  CogniPoint is designed for modern workplaces of all types — including open-plan offices, individual desks, meeting rooms, collaboration areas, reception zones, and large multi-purpose spaces. Because the sensor handles occupancy count, location, object detection, and traffic flow in one device, it is equally suited to fine-grained desk-level monitoring and broad area-level people counting. Any environment where understanding how space is actually being used translates into operational or cost benefits is a strong fit.