0
How many people came in, and at what times.
The gate count, by day and by hour. Put it next to the timesheet you received, and the gap is no longer a suspicion — it’s a figure you can raise at sign-off.
Vizor for general contractors and developers
The subcontractor billed for nineteen workers. How many came through the gate?
On a construction site, the number the money depends on is worker-days. It comes from the subcontractor’s timesheet and is taken on trust, because nobody stands at the gate with a notepad. Using your existing cameras, we count how many people come in and at what times, show the zones where people work without a hard hat or vest, and the time of the first and last movement on site. No new hardware. Facial recognition for the employees you register, with their knowledge; everyone else stays a silhouette.
General contractors, developers, project owners with several sitesCentru site · cam 1 · access gatelive simulation
entries today0
with PPE0
no hard hat0
on site now0
gate entryfull PPEno hard hat or vestmachinery, not counted
What Vizor sees on a construction site
The timesheet says how many were there. The gate says how many came in.
Between those two figures sits the subcontractor’s margin. You won’t find it in a spreadsheet, because the spreadsheet is made by the party doing the billing. It only shows at the gate, and the gate is already on camera at almost every site.
Worker-days, at the gate
pouring zone · 6 entries without a hard hat · 9–11
In which zones and hours people work without PPE.
Hard hats and vests are large patches of color, so they’re easy to see from a distance. The report gives the zone and the hour, not names: “in the concrete pouring zone, between 9 and 11, six entries without a hard hat.” Enough for a safety briefing, without pointing fingers.
When the real workday starts and ends.
The time of the first and last movement on site, by day. Plus the zones nobody passed through: either stalled work fronts or machinery sitting idle.
What we don’t doWe measure the gate, zones and hours — and, for registered employees, each person individually.
On construction sites, the first question we get is whether the system can show what each person is doing. The answer is yes, for the employees you register. But individual monitoring falls under Law No. 195/2024, under a strict regime: a documented legal basis, a data protection impact assessment (DPIA), informing every employee — and the employee’s consent is not a valid basis, because employees are in a subordinate position, so the basis has to be built on legitimate interest or on occupational safety obligations. We provide the tool and the documentation; the role of data controller and the responsibility remain yours. Tell us in the first conversation whether you’re going with named tracking, so we can prepare the right paperwork.
The site dashboard
Gate, zones, hours. Not named individuals.
The dashboard layout, as it looks in an installation. Switch sites and go through Dashboard, Alerts, Analytics and Reports. The figures are illustrative; in an installation they come from the cameras on site.
app.vizor.md · sites · dashboard
V
Vizor
On the phone
The gate count, straight to the site supervisor.
On a construction site, the decision that matters is made the same day: how many workers you confirm at sign-off, which work front is empty, where to send the safety briefing. The app is for that moment.
- The gate count, by day, next to the timesheet you received.
- The PPE alert arrives with the zone, the camera and the minute.
- The site report in the evening, without opening your laptop.
Example analyses · construction sites
Three scenarios, three decisions.
Our first real installations, in production, are in pharmacies in Moldova, where we count daily with a parallel manual check. On construction sites the metric is different: not customers and receipts, but people at the gate, zones and hours. The scenarios below show how to read the data, not what anyone has achieved.
⚠ Modeled scenariosThese are not results reported by clients. We measure your numbers in the pilot.
Modeled scenarioApartment building · general contractor
The timesheet claimed nineteen. The gate saw fourteen.
- Problem
- The structural subcontractor billed worker-days from its own timesheet. Nobody had anything to compare it with, and the discussion at sign-off ended with “that’s what the logbook says.”
- What it showed
- The gate count over thirty days: an average of 14 people a day, against 19 billed. The gap was consistent, not random, and it widened on Fridays.
- Decision
- The gate count went into the monthly sign-off record as a working document. The discussion is no longer about trust but about a list of days.
−26%worker-days confirmed vs. billed, over one month
people at the gate · by day30 days
Modeled scenarioIndustrial site · three work fronts
Hard hats were always missing in the same zone, at the same hour.
- Problem
- After an inspection, safety training was given to everyone, the same way, once a month. No visible effect, and nobody knew where the real problem was.
- What it showed
- Entries without a hard hat weren’t scattered: 8 out of 10 came from the concrete pouring zone, between 9 and 11, when concrete was being unloaded and work was rushed.
- Decision
- A short on-site briefing in that time window, plus a supply of hard hats at the entrance to the zone. Measured before and after, on the same cameras.
−62%entries without PPE in the target zone, over three weeks
no-hard-hat entries · by zoneone week
Modeled scenarioDeveloper · four sites
The workday started an hour later than reported.
- Problem
- The construction schedule was slipping on two of the four sites, with no explanation in the weekly reports.
- What it showed
- The first real movement on the two sites behind schedule came at 08:15 on average, versus 07:10 on the other two. An hour a day, for the entire project.
- Decision
- The start time went into the weekly report, next to physical progress. Sites are compared on the same measure, not on claims.
65 minaverage gap in start time, between sites
first movement per day · 4 sitestwo weeks
Our reference installation
The first location measured daily, with a parallel manual count and an automatic report every evening. The method is the same; on a construction site, all that changes is what we count and where.
How it works on a construction site
One day of installation. Without stopping work.
01
The cameras on site
The access gate is almost always on camera, for security. If a camera sees the entrance, we have what we need to count. For work fronts we add one or two ordinary cameras.
zero investment02
Gate and zones, drawn with you
We mark the entry line, the work zones and the materials area. Machinery and fixed posts are excluded from the count.
one day03
Figures by zone and hour
Images are processed on the Vizor server; the report gets figures: how many at the gate, how many entries without PPE per zone, the first and last movement. Per-person reports only for registered employees.
data minimization04
The count next to the timesheet
The gate count goes next to the timesheet you received, for the same days. That’s where the discussion with the subcontractor starts: with a document, not an impression.
monthlyPrivacy on site
We measure the gate, not the people who walk through it.
The system tracks silhouettes from frame to frame and produces figures by gate, zone and hour. It recognizes the faces of registered employees and can link their tracks, hours and zones to them; for anyone else, it stays at silhouettes and figures. The privacy notice for people entering the site is prepared together, as part of the pilot.
Privacy policy →Construction site pilot
One site. Three months. Numbers checked together.
We start with a single site, for the full length of one construction phase. In the first month we count manually on several days, at hours you choose, and compare the result with what the system produces. If the figures don’t match, we go no further.
New cameras
We work with your existing security cameras. Where coverage falls short, we add our own.
Work stoppage
Installation takes one day and doesn’t stop any work front.
Turnstiles or access cards
We don’t install access control. We count from the camera that already sees the gate.
Consent to track people
We don’t ask for it, because we don’t do that. We measure the gate, zones and hours.
Calculator · what the gap at the gate is worth
The calculation takes the billed worker-days and the percentage not confirmed at the gate. A mathematical estimate, not a promise: we measure your actual gap during the pilot, and on many sites it is zero.
Amount in question, per month20,800 MDL
unconfirmed worker-days32
per year249,600 MDL
Calculation: worker-days × unconfirmed percentage × rate × number of sites.
Read also
Related pages.
Facial recognition time & attendance
Employee attendance from your existing cameras, with no readers and no cards.
Read →Video cameras in the workplace
What is allowed, where it is forbidden, what the sign must say and how long footage is kept, under Law 195/2024.
Read →Warehouses and distribution centers
Dock occupancy, blocked aisles and the zones nobody walks into.
Read →Traffic studies for road design
Video-based counts you can re-check on the recording: volumes, vehicle mix, peak hour.
Read →Frequently asked questions
Questions from contractors.
Can you show what each worker is doing?
Yes, for the employees you register in the system. Each registered person is recognized by face, and their tracks are automatically labeled with their name: entry and exit times, the zones they worked in, entries without PPE. You do the registration, after informing the employee in advance — a facial template is biometric data. People who aren’t registered remain counted silhouettes, with no names.
How do you tell workers apart from drivers or visitors?
By path and duration. Someone who comes through the gate, goes to a work front and stays there is counted differently from someone who unloads a truck and leaves within ten minutes. The rules are calibrated to your site in the first month.
Does it work on our existing security cameras?
If a camera can see the access gate, we have what we need to count. For work fronts and hazard zones, we add one or two ordinary cameras where needed. We don’t install turnstiles and we don’t ask for access cards.
What happens in winter, in the rain or at dusk?
Counting at the gate holds up well, because people pass close to the camera. PPE detection drops in fog or darkness, and that shows in the report: low-visibility intervals are flagged, not reported as zero.
How accurate are the numbers?
In the first month we count manually on several days, at hours you choose, and put the two results side by side. The measured margin of error is published in the report. If you’re not satisfied, we go no further.
What does the law say about monitoring employees?
Individual monitoring of employees falls under Law No. 195/2024, in force since August 23, 2026, under a strict regime: a documented legal basis, a data protection impact assessment (DPIA), and informing every employee. Employee consent is not a valid basis, because employees are in a subordinate position: the basis is built on legitimate interest or on occupational safety obligations. We provide the tool and the documentation; the role of data controller and the responsibility remain yours. If you don’t want that regime, leave employees unregistered and stick to aggregate figures by zone and hour.
How does the count end up in talks with the subcontractor?
As a working document: the gate count, by day, next to the billed worker-days. We don’t decide who’s right; we put on the table a measurement both sides can check against the recording.
Contact
Show us a site. We’ll tell you what we can measure.
We don’t ask for an hour-long meeting. You pick a site, we look at the gate and the existing cameras, and we reply within 24 hours, in writing.
What happens next
Three steps, no pressure.
1We reply within 24 hours, in writing, with what can be measured at your location.
2Thirty minutes on a single location: together we look at what your existing cameras cover.
3If it makes sense for you, we propose a pilot on three locations. You can stop it at any time, with no penalty.