Asilla, Inc. (Headquarters: Machida City, Tokyo; Representative Director and CEO: Go Onoue; hereinafter "Asilla"), which pursues the vision of "a safe and comfortable world through the power of technology," is pleased to announce that it released Ver5.0, the latest version of its AI security system AI Security asilla (hereinafter "asilla"), on Wednesday, August 19, 2026.
Ver5.0 significantly broadens the range of security operations that asilla can support. By combining its core behavior recognition AI with AsillaVision, a VLM specialized for crime prevention, the system now narrows notifications to events that call for a response. Coverage has also been extended to exterior areas, capturing the facility as a whole across people and vehicles, indoors and outdoors. The update will be rolled out to existing customer facilities in stages.
The security industry faces a growing shortage of workers alongside an aging workforce. At the same time, the duties of security personnel extend across the entire facility: not only patrolling inside the building, but also checking the perimeter and parking areas, reviewing monitors in the disaster prevention center, and judging whether a notification warrants heading to the scene. Sustaining this broad scope of work with limited staff has become a significant challenge on the ground.
Since its founding in 2015, Asilla has focused on research and development in behavior recognition AI that interprets human activity from video. Built on that technology, asilla has been deployed at approximately 250 facilities in Japan and overseas since its release in 2022, accumulating a track record in security operations. Pursuing "AI that works in the field," Asilla has been evolving asilla from an AI security tool into a platform that supports security operations as a whole. Through roughly three years of operation, the company concluded that responding more deeply to on-site challenges required broadening the scope of what asilla can support.
In Ver5.0, the core behavior recognition AI is combined with AsillaVision, a VLM specialized for crime prevention, and, drawing on insights gained from three years of field operation, notifications are narrowed to events that call for a response. Coverage has also been extended to exterior areas, so parking lots, bicycle parking areas, and site perimeters can now be monitored continuously from the disaster prevention center in the same way as the building interior, enabling an early response when something unusual occurs.
This delivers both a reduction in the workload of security operations and an improvement in the quality of each individual notification.
For events detected by the behavior recognition AI, AsillaVision analyzes the situation including the surrounding circumstances.
By reading contextual information such as the condition of the person in the video and the surrounding environment, it determines whether the event warrants a notification.
Examples of use
AsillaVision is a domestically developed VLM specialized for the crime prevention domain, built by Asilla based on security camera video data accumulated at security sites (*1). Because it runs on a server inside the facility, it can be operated without transmitting video outside the premises. In early deployments, combining it with the behavior recognition AI has been confirmed to improve detection accuracy by up to approximately 80% compared with conventional methods (*2).
*1 Data is collected and used with the consent of the participating facilities and is subject to anonymization processing.
*2 Comparison against conventional methods based on operational data from early deployment sites and Asilla's own verification (subjects: falls and fights; verification period: March to May 2026)
Vehicle detection (entry and prolonged stopping) and bicycle detection (parking and prolonged abandonment), offered as an open beta since Ver4.4, are now available as official features.
Parking lots, bicycle parking areas, and site perimeters can now be monitored continuously from the disaster prevention center in the same way as the building interior, enabling an early response when something unusual occurs.
Examples of use
Asilla pursues "AI that works in the field" for security and develops systems that reflect the realities of security sites. This update aims to support the safety of the entire facility while reducing the day-to-day operational burden, by broadening the scope of what asilla can support. Asilla will continue to incorporate insights gained through field operation and build on updates in terms of both functionality and usability, contributing to safe and secure facility operations for its customers.
In applying AI at security sites, what we value most is that AI does not create a new burden for the people working there. However advanced the technology may be, if it increases the amount of checking and operation required and stands apart from everyday duties, it cannot be used continuously in the field.
What this requires is strong detection that captures anomalies accurately, together with a mechanism through which detected information connects naturally to daily security work. In Ver5.0, we combined behavior recognition AI with a VLM specialized for crime prevention to strengthen detection, and extended coverage from people inside the building to vehicles and bicycles in parking areas and along site perimeters.
In addition, by combining this with the already implemented integrations with intercoms, speakers, patrol lights, and shutters, detected anomalies can be shared with security personnel and connected to an initial response through facility equipment.
What we aim for is not to add work required to use AI, but a state in which AI operates naturally within everyday security duties and supports human judgment and response. We will continue to develop asilla together with our customers as a foundation that strengthens security across the entire facility without adding to the burden on the field.
AI Security asilla is a system in which AI analyzes existing security camera video 24 hours a day, 365 days a year, instantly detecting abnormal behavior such as violence, falls, and intrusion, as well as behavior requiring attention such as wandering, crowding, and signs of illness. Amid a deepening shortage of security personnel, it captures anomalies that are easily missed through human monitoring and immediately notifies security staff and facility managers. Because existing cameras can be used as they are, no capital investment in equipment is required, making it a next-generation security solution that maintains a high level of safety even with limited staff.
Representative: Go Onoue, Representative Director and CEO
Address: 1-4-2 Nakamachi, Machida City, Tokyo, Japan
Business: Development and provision of products and solutions based on behavior recognition AI
Official website: https://jp.asilla.com/
Asilla, Inc.
Public Relations: Nakamura
Email: pr@asilla.jp
.png)