Most cameras record. Some draw boxes. Pixonn watches every feed and describes, in plain language, what is actually going on, while it is going on. The AI does the watching; people supervise.
Video is the most underused asset an organization owns. It gets pulled up after something has gone wrong, and rarely while there is still time to do anything about it.
One operator, a wall of feeds, a long night. Human attention does not scale to a fleet of cameras.
Today's alerts are limited to perimeter protection and a set of pre-trained object types. Anything else goes by unnoticed.
Finding one moment means dragging through timelines by hand, so the operational and security insight in that footage stays locked up.
Take the cycle-store frame at the top of this page. Here is what each kind of system has to say about it.
The footage is there if someone reports a bicycle missing and has the time to go looking.
Both are permitted here. No rule matches, so nothing is raised. The tool in his hand isn't in its vocabulary.
Nobody anticipated this scenario or wrote a rule for it. The output is a sentence an operator can act on.
The same understanding of every scene drives both alerting and search. You describe what matters in a sentence instead of picking from a menu.
Alerts arrive as sentences, and you write the conditions yourself.
Search weeks of footage the way you search the internet.
An operator watches a wall of feeds while software flags a handful of pre-set events. Coverage is capped by how many screens one person can watch.
The system watches the whole fleet continuously and tells people what warrants their attention. It is an end-to-end system built for that, not an analytics layer bolted onto an existing one.
Pixonn pairs senior operating experience inside the video security industry with hands-on engineering for AI that has to run on real hardware, in the field, over bad connections.
Sean has spent his career leading engineering organizations, most recently as Head of Embedded Video AI at Motorola Solutions — one of the largest players in video security and access control. An entrepreneur and inventor, he leads Pixonn's strategy, partnerships and commercial work.
LinkedIn
Donovan is an embedded systems and edge-AI engineer who builds and maintains fleets of connected devices in the field — firmware, on-device machine learning and the unglamorous work of keeping hardware running over unreliable links. He leads Pixonn's technical work and authored the patent-pending approach the product is built on.
LinkedInWe're speaking with investors and with security teams who want to follow what we're building. Tell us a little about you and we'll come back to you quickly.
