What Security, IT and Loss Prevention teams need to know
For many multi-site businesses and organizations, adding AI to security cameras has traditionally meant one thing – rip all the existing cameras and recorders out and start from scratch with specialist equipment. Fortunately that’s no longer the only option, in many cases video analytics can now run on top of the cameras and recorders you already own.
This becomes more important if most of your sites aren’t all on the same setup, and in our experience onboarding multi-site customers, most aren’t. If you manage security across lots of locations, you’ve probably got older recorders at some locations and newer cameras at others – all of which have been installed by different people at different times. This can be a logistical nightmare when it comes to upgrading your systems.
Why you don’t have to replace everything
Take the example of one of our multi-store retail customers; they had cameras at some stores that were five years old, and brand new cameras at others – with no plans, or budget, to standardize their hardware anytime soon. Under the old model, adding features like people counting or license plate recognition meant new specialist cameras – because the intelligence lived inside the camera or recorder itself.
Platforms like Videoloft mean that no longer has to be the case. The camera can do what it does best – record footage and the AI e.g filtering out irrelevant motion, reading number plates, counting people etc can run separately. In terms of our customer – this meant that they could turn analytics on and off at individual locations, without touching their cameras or recorders.
What this means for your team
In our experience, a decision like this doesn’t normally sit with one person, there are usually a few teams involved.
IT want to know if this creates more work or risk than it saves. They usually have questions about bandwidth at sites with weaker connections, and how the cloud connection gets handled from a security perspective. A good system lets you add analytics per camera, based on what’s already set up there. At some sites, the cameras can connect straight to Videoloft, so there’s no recorder needed at all. At sites where you want to keep the recorder that’s already running, Videoloft can just back it up to the cloud and add AI on top, leaving the recorder in place. Either way, IT isn’t forced into the same setup at every site.
Loss Prevention doesn’t need to know how the AI works, just what it catches and how fast footage can be found. With Videoloft’s Smart Motion, you can draw a zone or a line around the area that actually matters, so alerts and searches only surface events that occurred there – instead of everything in the camera view. In an internal Videoloft comparison, a specific incident was found around ten times faster with Smart Motion than with standard motion detection. The standard search produced far more triggered clips to work through before reaching the same event.
Security leaders usually carry the hardest conversation, justifying the cost. The best way to do this is to pilot a handful of sites, prove it against the problem that matters most – shrinkage, how long incidents take to resolve, safety compliance etc – and then add more sites as the case holds up, rather than committing to everything at once.
What “ready for AI” actually means
Not every camera or recorder can run every feature. What actually matters is the camera’s resolution and position (especially for license plate recognition), how much bandwidth a site has, and whether the existing recorder or camera brand is supported. A realistic rollout checks sites against those three things first, rather than assuming every site is ready.
A simple way to roll this out
- Pick a handful of sites with a specific problem, think of sites with high shrink, frequent dispute or recent safety issues.
- Prove one or two analytics features work well to address the problem before adding more.
- Add sites one at a time as the results hold up.
Where this is heading
More multi-site businesses are moving this way – keep the cameras you have and add capability where and when it’s needed. Most teams we talk to have already decided AI is worth having – the key question is how they get there.
Want to see what this looks like on your own sites? Start a Videoloft pilot today.
Frequently asked questions
Can AI analytics be added to any existing security camera?
No, firstly you’d need to check if your cameras are compatible with the analytics platform you want to use. Also, you need to think about camera placement – position, lighting and connectivity can all affect what is possible.
Can AI work with analog cameras?
In lots of cases they can. Analog cameras connected to a compatible DVR can potentially be brought into a cloud or analytics platform through the recorder or an adapter. Support will depend on the recorder, camera setup and chosen platform.
Can different camera brands be managed from one system?
Yes, Video Management Systems like Videoloft can support cameras and recorders from multiple manufacturers.
How much does it cost to add AI to existing cameras?
Costs vary according to the platform, number of cameras, storage period and analytics selected. A per-camera model can allow businesses to test analytics at a limited number of locations before deciding whether to expand them.
