When the inspector arrives, what can you actually prove?
A fatality at a mine starts an investigation that can run two years. The question that decides the cost is whether you can prove what was happening before it did. Without video-verified work, you are reconstructing the last shift from memory. With Stratos, you are reading it off the record.
Total cost
$4.24M
$8.88M
Production days lost
2.9
6.2
Records available
12/12
8/12
Exposure resolved
day 400
day 690
with Stratos · without
1.2 months
cut from a serious safety incident while on project, because the video-verified data already exists when the inspector arrives.
Investigation compression model
One incident, two clocks.
The same fatality at the same mine. The cyan clock reads off a record that already existed. The red clock reconstructs the same event from memory. The distance between them is the cost of not having video-verified work.
Cumulative cost
$0·$0
difference $0
Production days lost
0.0·0.0
full-shift equiv.
Milestones reached
0/8·0/8
of 8
Records available
12/12·0/12
none on the left
Statutory notices filed
$1.20M delay cost
Document requirement satisfied
$3.86M delay cost
Scene released
$1.17M delay cost
Full production resumed
$1.55M delay cost
Root cause established
$1.12M delay cost
Orders closed
$1.54M delay cost
Regulatory exposure resolved
$1.19M delay cost
Inquest concluded
$720k delay cost
Stratos enabled · 24 months
$4.24M
total cost of the incident, platform included
2.9 production days lost · 12 of 12 records at T+0 · resolved day 400
Without · 24 months
$8.88M
total cost of the incident
6.2 production days lost · 4 of 12 records missing · charges day 690
Illustrative model, not a claim about any operation. Same incident, same mine.
The evidence board
The instant it happens, what can you actually audit?
An inspector arrives the same day with the power to take and copy any record. These are the twelve things the investigation will ask for. Each is a record an Ontario mine is already required to keep, produce, or explain the absence of.
Without video-verified work
8/12
arrive eventually, partial and contested. Four are absent.
With video-verified work
12/12
available right now, in one export, from T+0.
The 24 hours before the event, in this heading
A contested reconstruction assembled from 200+ interview hours. Two accounts of the last hour do not agree.
Continuous and time-stamped: activity, persons, equipment and controls, 24 h before and 1 h after.
The supervisor's examination of this face, this shift
A ticked box in the shift logbook, found on day 21. It records that an examination happened, not what was seen.
07:05, R. Tremblay, four minutes at the face, and what was in view while he was there.
The end-of-shift danger record and its countersign
The countersign is in the book. Whether the incoming crew was actually advised is a matter of recollection.
Open hazards at shift end, the countersign, and the pre-shift briefing where the crew was advised.
Ground support installed to pattern before the advance
The heading was mucked and re-supported to reopen the level. The pre-advance state no longer exists anywhere.
18 of 18 bolts to pattern, screen overlap confirmed, 06:14 to 06:42, camera 1435-07.
Whether scaling was completed before bolting
Two crew members remember scaling. Neither can place it in time. The Crown will test that, and so will the inquest.
Scaling observed and timed before the first bolt, in the same clip.
Who was present, and where they were standing
The roster gives names. Positions at the moment of the event come from three accounts that conflict.
Presence and position for every person in the heading, second by second.
Training and accreditation of those present
Modular training records pulled from the training office. This one exists either way.
The same records, joined to observed practice so the Signing Authority sees method against training.
Equipment movements and interactions nearby
Dispatch logs place a machine on a level, not in a heading, and record no interactions at all.
Every vehicle and pedestrian interaction in the area for the shift.
Energy isolation state of nearby equipment
The lockout register, if it was filled in. Whether the lock went on before entry is recorded nowhere.
Lock applied 05:58, before entry, with the verification step performed.
The procedure actually being followed at 06:40
The written procedure is on file. What the crew was doing is now a question for a jury.
Observed practice against the written method, including the two steps taken out of sequence.
Ground instability in the preceding shifts
The s.72 register, where it was kept. Exists either way.
The same register, plus the observed ground behaviour behind each entry.
Chain of custody for all of the above
There is no chain of custody on a recollection. Provenance gets argued, not proved.
Hash on capture, export log, access history, retention hold applied at T+0.
What the compression is worth
Every milestone moves. The dates stop being estimates.
| Milestone | Without | With Stratos | Compression | Why the date moves |
|---|---|---|---|---|
| Scene released | day 3.0 | day 1.5 | −1.5 d | The pre-incident state is held independently of the rock, so the inspector reviews conditions before the event instead of reconstructing them from it. |
| Statutory notices filed | day 2.0 | day 0.5 | −1.5 d | Prescribed content is generated from the record. Counsel edits facts instead of assembling them, and the report is not revised twice in the following fortnight. |
| Full production resumed | day 12 | day 5.5 | −6.5 d | Return to work turns on demonstrating the control was in place and is verified now. A dated restart can be published on day two. |
| Document requirement satisfied | day 21 | day 1 | −20 d | One export with hashes and an access log replaces a search across four departments and two companies. This is where failure-to-produce orders get written. |
| Root cause established | day 60 | day 17 | −43 d | The objective timeline is the first artefact rather than the output. Interviews explain the record instead of building it. |
| Orders closed | day 90 | day 21 | −69 d | Fewer orders, and compliance with a practice-based order is demonstrable from the day it is issued, so the worker member has something to sign. |
| Inquest concluded | day 690 | day 540 | −150 d | Mandatory for a mine death either way. A verified timeline is an exhibit rather than three weeks of public reconstruction from testimony. |
| Regulatory exposure resolved | day 690 | day 400 | −290 d | Due-diligence evidence delivered in week one narrows what the Crown can prove and opens the administrative route, which bars prosecution on the same contravention once paid. |
The record that decays
Without video evidence, you have no evidence it got done.
Three of the twelve records an investigation demands are states of the world at a moment. They are not files you can pull later. Once the heading is mucked and re-supported to reopen the level, they are gone. No file holds them. No interview recovers them.
The best anyone can offer is a recollection. The Crown will test that recollection. The family will hear it. A court will decide due diligence on the memory of a crew that just lost a colleague.
gone with the heading
Pre-advance support state
gone with the heading
Whether scaling was completed before bolting
gone with the heading
The procedure the crew was actually following at 06:40
There is no chain of custody on a memory. Provenance gets argued, not proved.
Video-verified work is the only record that exists before the ground moves. Everything else is reconstructed after, from people who were there, under the worst pressure of their lives.
The camera is the witness that does not forget, does not leave, and does not need to be asked.
The scheduling story
The cost is the days you cannot plan.
An unbounded investigation does its damage before the days are lost. Crews stand down without a return date. The development contractor demobilises rather than wait. The two-week outlook meeting produces a plan that is void by Thursday. The quarterly target quietly becomes unreachable without anyone deciding that it has. With Stratos, the whole sequence is datable within the first day, because the record already exists.
Share price vs. day before
The gap is the uncertainty premium
- Without
- With Stratos
The same incident. One operator ends the two years below where it started. The other is back above. The market prices the uncertainty, and Stratos removes it.
$4.24M
$8.88M
total cost of the incident
with · without
2.9
6.2
production days lost, full-shift equivalent
with · without
12/12
8/12
records available, 4 missing without
with · without
day 400
day 690
regulatory exposure resolved
with · without
The pilot pays for itself on halt days alone. One verified record at T+0 is worth more than two years of reconstruction. Book the demo and run the model on your operation.