We are taking on a small number of pilot sites. You get a Hive on your cameras, direct access to the two people who build this, and real influence over which detection classes ship next. We get the one thing we cannot buy — a real building, with real traffic, telling us where the model is wrong.
A Hive sized to your camera count, configured remotely against your actual site. The two of us on the other end of a phone rather than a ticket queue. Detection classes prioritized around what is actually going wrong in your building. And whatever commercial terms we agree for the pilot, held for the pilot period.
A real site. Every deployment so far has taught us something a test bench could not — a reflective floor that reads as motion, a delivery route that looks like an intrusion, an angle nobody thought to check. That feedback is the entire reason the program exists, and it is worth more to us than the revenue.
We would rather turn down a site than take one we cannot help. These are the things that make a pilot work, and the things that make it fail.
Four or more IP cameras producing RTSP or ONVIF, pointed at things that matter, on a network with a spare switch port. If the angles are poor, no software fixes that — and we will tell you before you commit.
"Stock walks out of aisle three." "Falls in the back room get found twenty minutes late." A specific, recurring incident beats a general wish to be more secure, because it gives us something measurable to aim at.
The pilot lives or dies on feedback. We need one named person who will tell us when an alert was wrong, ideally within a day. Sites where nobody has time for that get no value from being early.
You are coming in before the polish. Expect false positives in the first weeks, expect to adjust thresholds, expect the occasional feature to arrive mid-pilot. If you need something finished, buy it later at list price and we will not be offended.
Sites needing a compliance certification we do not hold yet. Sites where the only classes that matter are on the roadmap rather than shipping. Anywhere the cameras would need replacing first — that is a different project, and not ours.
If you run several locations, we would rather pilot one properly than all of them badly. Pick the site with the worst recurring problem, prove it there, then roll out with numbers behind you.
Make, model, count, and footage from the angle you have least faith in. We run it through a Hive and come back with a written read: what we would flag, what we would miss, and which of your angles we would not put a promise on. If the answer is that we cannot help, that is where it ends.
Scope, duration, which detection classes are in, what happens at the end, and the commercial terms. Written down before anything ships, so nobody is guessing later.
One free port on your switch and a power outlet, roughly ten minutes, no specialist needed. We do not send an engineer. From the moment it dials out we configure it remotely — discovery, zones, schedules, rules and escalation.
The first weeks are where a site finds its own false-positive personality. Every dwell time, zone and confidence cut-off is in your console to adjust, and we are on the other end when you want a second opinion on a setting.
What fired, what was real, what it caught that would otherwise have been found the next morning. That review decides whether the pilot becomes a normal deployment, and it is also the honest place to conclude it did not work for you.
Pilot or not, the pricing shape is the same two parts — a one-time setup fee and a monthly subscription per site. See how pricing works.
A pilot is not a free trial with the support switched off. It is a working relationship with a short list of obligations on our side and a short list on yours.
That last one is the only marketing ask in the program, and it is genuinely optional.
Sample data shown.
Make, model and count, plus footage from the angle you trust least. You get a written read on what we would flag and what we would miss — before either of us commits to anything.
