EyeKnow

EyeKnow AI develops a computer-vision and deep-learning platform that connects to standard, off-the-shelf security cameras and CCTV/DVR systems and analyzes footage in real time to identify aggression and violence in care environments. The system is designed to detect subtle physical interactions between people rather than only overt violence, including strong arm pulls, forcefully seating a person and violently feeding someone, which the company argues existing violence-detection systems cannot resolve at the required sensitivity. When an aggressive action is detected, the platform captures a short video clip of the moment and pushes it through a mobile app or dashboard to a designated stakeholder such as a chief clinical officer, security guard or family guardian, so the incident can be reviewed and acted upon.

The intended settings span child daycare centers and nurseries, at-home and institutional elder care, nursing homes and geriatric centers, autism care facilities, and hospitals and psychiatric centers, where the company notes aggression is frequently directed from patients toward healthcare workers. The company also cites potential future use in public-space violence prevention for municipalities, schools, prisons and police stations. Aggression is framed as multi-directional: between care receivers, from caregiver to care receiver, or from care receiver to caregiver.

The company frames the problem using Israeli abuse cases and WHO figures it cites, including an estimate that one in five elderly people in Israel experience some form of violence by caregivers and that 15-30% of seniors in nursing homes or home care experience verbal or physical violence. It sizes its addressable market at roughly 113 million seniors receiving professional care worldwide, about 81 million of whom live at home, plus childcare markets that in Israel include roughly 2,100 daycare centers and 3,600 nurseries.

Founding story

Tomer Sagi, a parent with more than ten years of experience in the technology industry who previously worked as a senior project manager for Israel's Ministry of Social Affairs starting in 2008, founded the company after identifying the need to protect people who cannot protect or report on their own behalf. The company's public materials tie its formation to high-profile Israeli abuse cases, including the death of Yasmin Vinta and the Carmel Mauda childcare abuse convictions, as well as reported abuse in nursing homes.

Business model

EyeKnow sells a monitoring service to operators of care settings. Engagements begin with a two-to-three-month pilot in which the company's cloud systems connect to the client's DVR/CCTV equipment over the internet and the detection algorithms are tuned to the specific site; after the pilot the company deploys an on-premise solution, including a computer workstation installed at the customer location, which it positions as necessary for privacy protection, security and low latency. Alongside detection and alerting, the platform provides facility owners with aggregate statistics and trend insight into aggressive behavior, which they can act on through measures such as additional staff training and stress-management instruction.

Customer engagements start with a two-to-three-month cloud-connected pilot and move to an on-premise deployment that includes an installed workstation at the customer site; the sources do not state pricing terms.

Traction

In early pilots, a daycare owner was alerted to a caregiver forcefully pulling a child's arm, and at an Israeli nursing home the chief nurse received alerts about a caregiver behaving aggressively toward a resident, after which a committee reviewed the incident and imposed disciplinary action. The company reported being in talks with one of the top 10 U.S. health organizations. A survey it commissioned with Geocartographia found about 40% of respondents considered the service relevant to them, rising to 47% among those with a close elderly relative or aged over 70, while 71% viewed the technology as innovative and unique versus existing solutions.

Latest developments

The company reported developing a new proprietary algorithm set intended to deliver more accurate predictions across a broader range of incident types without relying on the large labeled datasets that supervised models require, and said it was in discussions with one of the top 10 U.S. health organizations about integrating its monitoring technology.