Blog
.webp)
By Ed Burnett, Chief Security and Operations Officer, Cloudastructure
Disclosure: I run security and operations at Cloudastructure, one of the companies compared below. I have tried to describe every competitor accurately and to say plainly what each one does well. Where the comparison grid comes from a customer's own evaluation, I have said so, and where I could not verify something I have marked it rather than guessed.
Short answer: The truck yard surveillance market has consolidated around six kinds of provider — Cloudastructure, Evolon, AMAROK, ECAM (GardaWorld's merged Stealth Monitoring and ECAMSECURE business), Birdseye, and traditional central monitoring stations. The only meaningful comparison between them is how their AI works, because nobody is buying passive cameras anymore and nobody can afford enough human eyes to replace them. The two questions that separate these companies are whether the AI is real, and whether you can keep the cameras you already own.
A decade ago, comparing yard surveillance vendors meant comparing camera specs and storage. That comparison is over. Sensors, lenses, and weatherproof housings are commodities, and every serious provider in this market now leads with analytics.
Two older models still get sold, and both fail for the same reason — neither one stops anything.
Cameras alone produce footage. Footage tells you what you lost after the trailer is on the interstate. With Verisk CargoNet reporting roughly $725 million in cargo theft losses in 2025 and an average reported loss of $564,009 per incident in Q2 2026, "we have it on video" has stopped being an acceptable outcome.
Live human monitoring without AI is the expensive version of the same failure. A human operator loses reliable attention within minutes watching a wall of static feeds — this is well-documented in vigilance research, and it is why guard-based monitoring misses events. Worse, the cost model is linear: every camera you add and every site you open needs proportionally more operator time. AI inverts that. The computer watches everything continuously and never gets bored; the human is called in only when there is something real to judge.
So the useful question is not "who monitors my yard." It is what does the AI actually do before a human sees anything, and what does that architecture cost me over five years.
Five questions sort this field faster than any feature grid:
Model: Cloud-based AI surveillance plus in-house remote guarding, camera-agnostic, month to month.
Cloudastructure runs analytics in the cloud rather than on a chip inside the camera, which means the models improve for every customer at once instead of being frozen at the moment of purchase. The platform connects to roughly 99% of commercial-grade IP cameras already in the field, with analog cameras joining through an encoder — so most yards keep the hardware they have.
For trucking specifically, that includes agnostic license plate, DOT, MC, and trailer number reading, AI search across historical footage, rover patrols in addition to event response, talk-down through yard speakers, and monthly per-lot reporting. Monitoring is in-house — Cloudastructure employs its own operators rather than subcontracting to a third-party central station. This is the part I own, so I will be specific about how it works. I train teams myself, and they are trained to the specific site they cover rather than to a generic playbook. My background is 30 years in logistics security, most recently as VP of U.S. Security and Global Fraud Investigations at UPS, where my remit covered more than 376,000 colleagues; the failure modes in a freight yard are not theoretical to me. There is no annual contract, and the company is SOC 2 certified.
The part I would ask about if I were buying: during slow periods, our operators perform machine learning markup on the footage from their own sites — labeling and correcting what the model saw. The AI watching your yard therefore gets measurably better at your yard: your gate geometry, your traffic patterns, your normal, your anomalies. Most providers treat operator downtime as idle cost. We treat it as training data.
That bends the cost curve the right way. In a labor-first model, every yard you add costs proportionally more operator time, forever. In ours, the operator hours you buy this quarter reduce the operator hours your sites need next quarter, because the model absorbs the judgment. And because one company owns the cameras' data, the AI, the operators, and the escalation, there is exactly one phone number to call when something is missed — nobody can point at a subcontractor.
The honest caveat: a cloud platform needs solid internet at the site. For remote lots with poor connectivity, that has to be solved during design rather than discovered after install.
The result that matters: one trucking customer has not suffered a single cargo theft across any of its lots since Cloudastructure took over monitoring in December 2025 — nine consecutive months and counting, in a period when national cargo theft losses were setting records. That is the only benchmark worth quoting in this category: not detections logged, not clips reviewed, but thefts that did not happen. Across all deployments, live intervention deters roughly 98% of detected incidents before anything is taken.
Model: Cloud-based AI video monitoring, camera-agnostic, no long-term contract.
Evolon is a Dallas-based AI monitoring provider with a genuinely plug-and-play deployment. Its AI search and camera-agnostic architecture put it closest to Cloudastructure on paper.
Trade-offs to check: Evolon is a newer firm and several capabilities customers ask about in trucking — talk-down, license plate and trailer number reading, rover patrols, in-house monitoring — were either absent or still maturing in a recent side-by-side customer evaluation. In April 2026, AMAROK acquired a 45% stake in Evolon and placed two executives on its board. That is a strategic investment rather than a full acquisition, but it does tie Evolon's roadmap to AMAROK's perimeter-hardware business. If you are evaluating Evolon standalone, ask directly how the AMAROK relationship affects pricing, packaging, and product priorities going forward.
Model: Leased electric perimeter fencing with video monitoring, hardware-supplied.
AMAROK's differentiator is real and it is physical: an electrified perimeter fence that stops entry rather than recording it. For certain yards — high-value, fixed footprint, chronic intrusion — that is a legitimate answer, and no software-only vendor should pretend otherwise.
Trade-offs to check: it is a lease model built around AMAROK's own hardware, which in a recent customer evaluation meant replacing all existing cameras — the reason that customer marked it cost-prohibitive. It is not camera-agnostic and not cloud-based in the way the software-first vendors are. Fencing also does nothing about tailgating, internal theft, driver-facilitated loss, or anything that walks through your gate with a legitimate-looking manifest.
Model: Large-scale live video monitoring, fixed and mobile units.
ECAM is the biggest footprint in North America: GardaWorld completed its acquisition of Stealth Monitoring in October 2024, merged it with ECAMSECURE, and launched ECAM in April 2025 with roughly 140,000 cameras under monitoring. Scale, national coverage, and mobile surveillance units are genuine strengths, and the combined business is investing in AI.
Trade-offs to check: the customer evaluation behind this article flagged missed events, and too much human element for not enough AI. That is one customer's experience at one point in time and the company is actively investing in AI — but it is exactly the failure pattern that human-first monitoring produces at scale, and it is worth pressure-testing in a trial rather than taking on reputation.
Model: "Maximum Telepresence" — AI detection plus trained remote agents, hardware supplied and installed.
Birdseye is the most visible brand in trucking-specific yard surveillance and deserves the attention. Founded in 2011 and headquartered in Mississauga, Ontario, it has built genuinely operations-aware products: GateCore handles driver verification, credential checks, BOL validation and seal confirmation at the gate; YardCore tracks trailer movement, dwell time and staging; SafeCore covers perimeter and after-hours. For yards where the gate process itself is the problem — not just theft — that operational depth is a real advantage, and few competitors match it.
Trade-offs to check (all from Birdseye's own published material rather than third-party criticism):
Model: Human operators, your equipment, your integrator.
Traditional CMS providers give you a direct relationship with the monitoring team, which some operators genuinely prefer. But monitoring-only means you still need a separate technology vendor for cameras, storage, and analytics, and the two relationships have to be managed — and blamed — separately when something is missed.
The grid below is adapted from an evaluation performed by a Cloudastructure trucking customer comparing vendors for its own lots. It reflects that buyer's findings at the time of evaluation, not verified vendor statements. Birdseye entries are drawn from Birdseye's published material. Verify anything decision-critical directly with each vendor.
Biggest plus, by vendor: Cloudastructure — one-stop platform with in-house monitoring and no annual contract. Evolon — plug-and-play deployment. AMAROK — physically stops entry with electric fencing. ECAM — largest monitored footprint in North America. Birdseye — deepest gate and yard operations tooling. CMS — direct relationship with the monitoring team.
Main caveat, by vendor: Cloudastructure — needs solid internet at the site. Evolon — newer firm, several capabilities still maturing, now 45% owned by AMAROK. AMAROK — lease model requiring all-new cameras. ECAM — human-heavy with reported missed events. Birdseye — labor-intensive model with supplied hardware. CMS — monitoring only; requires a separate technology vendor.
Every vendor on this list says AI. Three tests separate the claims:
Run a trial in your worst yard, not your best. A windy night with plastic bags, headlights sweeping the fence, and a raccoon problem is the real benchmark. Ask for the false-alert rate from that week, not from a demo reel.
Ask what happens between detection and dispatch. Time the sequence: AI flags an event, operator verifies, voice-down goes out. If nobody can tell you the average, nobody is measuring it.
Ask where the intelligence lives. On-camera AI chips are frozen at purchase — when the model needs upgrading, the fix is a truck roll and a hardware swap on every pole. Cloud analytics retrain centrally and every site benefits overnight. This is the single biggest determinant of what your system will be worth in three years.
Every company on this list will show you AI in a demo. The differences that matter are structural, and they show up in year three, not week one: who owns the cameras, where the intelligence lives, whether the model improves without a truck roll, and whether a human is verifying AI-flagged events or staring at a video wall hoping to catch something.
Cameras record. Human-only monitoring costs too much to scale and misses things anyway. AI does the watching, people do the judging, and the cloud does the improving — that combination is the only one that gets cheaper and better at the same time.
Cloudastructure provides cloud-based AI surveillance and in-house remote guarding for freight yards, distribution centers, and truck stops — real-time detection across large perimeters, live talk-down, verified dispatch, agnostic plate and trailer reading, and analytics that keep improving. It works with roughly 99% of the cameras you already own, runs month to month with no annual contract, and is SOC 2 certified.
Book a 30-minute assessment on your own lots, watch it stop a theft in real time, or call 1 (650) 644-4160.
.png)
Ed Burnett is Chief Security and Operations Officer at Cloudastructure, where he leads the remote guarding organization and trains its monitoring teams. He previously served as VP of U.S. Security and Global Fraud Investigations at UPS, responsible for protecting more than 376,000 colleagues and securing logistics operations across the United States, with a career spanning risk management, threat mitigation, and fraud investigation at global scale.
Sources: Verisk CargoNet cargo theft data (2025 annual, Q2 2026); GardaWorld acquisition and ECAM launch announcements (October 2024, April 2025); AMAROK–Evolon investment announcement (April 2026); Birdseye Security Solutions published company materials. Vendor comparison grid adapted from a customer-conducted evaluation.
.svg)
.svg)
.webp)
.webp)
An intimate sunset gathering for property management leaders and industry partners. Space limited to 75 guests.
RSVP Now →