Loading…
Loading…
The scenario: a large technology company runs a 12-building campus with more than 8,000 employees and 1,500 cameras on a fragmented system. Each building has its own recorder from a different era and vendor, so security staff walk between control rooms to review footage. There are no analytics, visitor handling is manual, unauthorised access keeps recurring, and investigations take days.
The solution is designed to bring all 1,500 cameras onto Visylix without replacing them, with edge servers in each building and a central security operations centre (SOC) on the company's own infrastructure. Visylix will check tailgating against badge swipes from the access-control system, raise watchlist and unknown faces in restricted areas for an operator to confirm, and let investigators follow plate and person journeys across the campus in seconds.
A campus like this grows building by building, and each addition brings its own recorder. The result is a patchwork of systems from as many as eight manufacturers, with separate logins and no shared view.
When someone suspicious is spotted in one building, nobody can quickly check where else they went. Investigations that should take minutes take days of manual scrubbing, and repeated unauthorised access eventually forces the decision: one platform, with AI, tied to the access-control system.
The migration is phased so coverage never drops. Visylix edge servers will go into each building and connect the existing cameras over RTSP and ONVIF, unifying all 1,500 in one interface. Native archive access for Hikvision, Dahua, Uniview, I8H, Bosch and Panasonic recorders keeps past footage reachable throughout.
Visylix will connect to the access-control system (Brivo, LenelS2, HID, AMAG, Genea/Openpath or a generic connector) and check every door event against the people actually passing, alerting only when bodies exceed badges. Watchlist faces and unknown people in restricted areas will be raised for an operator to confirm.
The SOC will run floor-plan maps of every building with live alarm badges, an alarm and incident console and video walls up to 6x6. Investigators will ask Radha, the on-premise AI copilot, to find a person or vehicle in plain language. Plate journeys are exact, and appearance-based person journeys surface candidate routes across cameras for the investigator to review.
Staff will sign in with Active Directory, LDAP or passkeys, and 46 permissions across 4 roles will decide who sees which buildings and functions. Every view, export and Radha action is written to a tamper-evident audit trail, and evidence is exported as a self-contained player sealed with SHA-256.
Each building is planned to transition over a single weekend, with on-site training so the team is confident before it goes live.
The solution is designed to bring every camera onto one platform in 10 weeks, to raise a tailgating alert within five seconds of the door event, and to turn investigations that took days into searches that return in under 30 seconds, freeing the security team for proactive work.
The same design scales to a multi-country campus estate, with Aptibit as the implementation partner.
Our solution architects will design a deployment tailored to your industry, scale and integration requirements. Let us scope this blueprint to your estate.