KSIistheevolution of people counters.
Other vendors sell hardware that becomes obsolete, analytics limited to the model purchased, and require hardware installations per location that multiply costs with every new location. KSI operates on a completely different model.
Software-first, not hardware-first
THE BROKEN MODEL
The industry model is to sell hardware. Sensors, proprietary cameras, counting devices. Every new capability - heat maps, demographics, re-identification - requires buying the more expensive model. When AI advances, the hardware becomes obsolete. The vendor's business depends on the customer continuing to buy.
THE KSI DIFFERENCE
KSI is pure software. It runs on any existing camera - IP, analogue, DVR, NVR, any brand. Every AI advance is deployed as a software update, without touching the hardware, without additional investment, without planned obsolescence.
Like Google Maps vs. a dedicated GPS. The GPS offers what it had when it left the factory. Google Maps works on the phone that already exists and always has the latest version, the latest data, the latest intelligence. KSI works the same way: the cameras are the hardware, KSI is the Google Maps running on top.
Anonymous re-identification - The superpower no competitor has
Counting how many people enter a door is trivial. Any sensor can do it.
What no traditional sensor can answer is: did the person who just arrived at the checkout already visit the tools section? Has the visitor who has been in the store for 40 minutes browsed three categories or stayed in just one?
KSI assigns a unique, anonymous ID to each person - without storing images, without biometric data, without facial recognition. Just a morphological signature that allows reconstructing the complete journey of each visitor.
The same person. Two cameras. One ID. No personal data stored.
more context per visit
In retail
Reconstructs the customer's complete journey through the store. Which zones they visited, in what order, where they spent the most time, where they dropped off. Without Re-ID, there are only isolated counts per camera.
visitor journey
In shopping malls
Understands the real behaviour of visitors throughout the entire mall. Which stores they visited, how long they spent in each, which routes are most frequent.
times per passenger
In airports
Measures the real time of each passenger through the process: check-in, security, waiting area, boarding. Identifies real bottlenecks, not averages that hide peaks.
Why anonymous Re-ID is better than biometric recognition
Some solutions offer re-identification through facial or biometric recognition. The problem is threefold: it requires very costly specialised cameras that don't scale, demands explicit consent from each visitor for commercial purposes, and is under growing regulatory scrutiny. KSI operates on morphological signature - no face, no biometric data, no consent required. Compliant with GDPR, CCPA and the world's strictest privacy standards.
Centralised architecture - Deployable across thousands of locations without multiplying costs
vs. per-store server architectures. With 100 stores, you pay the same infrastructure cost as 10 stores with the traditional model.
THE PROBLEM
Most video analytics platforms require installing a dedicated server in each store. With 10 stores, you pay for 10 servers. With 100, you pay for 100. Infrastructure costs grow linearly with every new location - making real-scale deployment unviable.
THE KSI SOLUTION
KSI is designed from day one for global chains. Centralised architecture that allows deployment across thousands of stores without multiplying infrastructure cost. A single platform, granular permissions per store and region.
- →No server per store
- →Centralised management from a single platform
- →Granular permissions and roles by location
- →Comparisons between locations, regions and formats
The operations team does not have to worry about the data. That is KSI's responsibility.
Working on existing cameras does not mean data quality is the customer's responsibility. The system continuously monitors the connection, the status of each camera and the generation of analytics in real time.
Has it ever been discovered weeks later that a sensor had failed and no one had detected it? Or were there doubts about whether traffic dropped because it actually dropped, or because a camera stopped working ten days ago? With KSI that question does not exist.
24/7 monitoring
Automated system of continuous connection monitoring, camera status and analytics generation. Problems are detected before they affect the data.
KSI knows first
When a disconnection occurs, the KSI team receives the alert before the customer's team. Follow-up continues until complete resolution - without the operations team having to manage anything.
The data is always complete
In case of interruption, statistical backfill based on historical patterns is applied so the data series has no gaps. Trend analysis is not broken by a one-off technical issue.
The question "did traffic drop or did the camera fail?" disappears from operations for ever.
How does KSI compare to traditional counters and native camera analytics, feature by feature?
See KSI vs. other solutions →