Built from real operations
ATLASKAI builds KAI, the AI operating layer for physical security: software that senses, understands and acts across the systems security teams already run.
ATLASKAI was founded by three people who have spent their careers building intelligence platforms and turning them into global businesses.
That work led to a simple conclusion: the bottleneck in physical security is no longer sensing. Everything is already recording. What no team has is a single, current picture of what it all collectively says, and the tooling around it still assumes a human will assemble that picture by hand.
KAI is our answer: an intelligence layer that reads every connected system together, keeps the full timeline, and drives the response through the systems already on site.
Yair Pecht
President & Co-Founder
- Four-time CEO across cyber-intelligence, telecom and IT
- Co-founder & CEO of Interionet, a profitable cyber-intelligence business
- Former CEO of NSO Group and of Alcatel-Lucent Israel (now Nokia)
Theo Kletz
CEO & Co-Founder
- Major with 11 years of experience in the IDF
- Former platoon commander in the SOD
- Led large-scale technological long term operational tasks
Sharon Oknin
CTO & Co-Founder
- Two decades building deep-technology products against sensor systems
- Co-founder & CTO of Interionet, leading the cyber-IoT platform end to end
- Architect of KAI's multi-agent platform and the CHRONOS engine
BG. (Res.) Asaf Kochan
Former Commander, Unit 8200
Co-founder & President of Sentra, after a 31-year career leading one of the world's foremost technical intelligence agencies.
BG. (Res.) Eyal Zelinger
Former IDF Chief Signal Officer
Thirty years commanding the IDF C5I Branch, later Global Defense VP & GM at Gilat Satellite Networks.
Context beats detection
A detection is only useful if the system knows what came before it and what should happen next.
Fit the field
Operations teams cannot pause to replace hardware. Software has to adapt to what is installed.
Prove it on site
We connect to a real environment, measure the change in alert quality and response time, then expand.
Talk to the team
We are happy to walk through the architecture, the deployment model and what a pilot looks like.
Get in touch