The gap between an inspection finishing and the report arriving is dead time, and almost nobody prices it. It sits in the project programme as an assumption — three days, maybe five — and everything downstream of the hold point waits inside it.
We built our reporting to close that gap, and the commercial case turns out to be much larger than the technical one. Here is where the time actually goes, and what changes when results are available while the technician is still on site.
Where the days go
Traditional inspection reporting has a shape that has barely changed in thirty years. The technician works on site and records results on paper. The paperwork travels back to an office. Someone types it into a template. A Level 2 or Level 3 reviews and signs. It is converted to PDF and emailed.
Each step is short. The queue between them is not. The paperwork waits for the van to come back. The typing waits for someone to be free. The review waits for the reviewer to be off site. Three days is not three days of work — it is a few hours of work spread across three days of waiting.
And the transcription step introduces the errors. A number read off a gauge, written on a damp sheet in a compound, typed by someone who was not there, is a number with three opportunities to go wrong before anybody senior looks at it.
What the delay actually costs
Four things, in rough order of expense.
Standing time at a hold point. If a weld cannot be painted, insulated, or built over until the NDT result is in, everything behind it queues. On a fabrication shop with a sequenced programme, three days on a hold point is three days on everything downstream of it. This is the big one, and it is almost never attributed to the inspection.
Rework discovered late. A defect found on Monday and reported on Thursday is a defect that has had three days to be built over. The excavation is deeper, the access is worse, and sometimes the coating is already on. The cost of a repair rises with how long it stays undiscovered, and it rises fast.
Remobilisation. If a result requires a re-shoot or a complementary method, and the crew has left, that is a second mobilisation. On a site half a day away, that is a day of travel to do an hour of work.
Decisions made on assumption. The most expensive one and the hardest to see. When results are slow, people stop waiting for them. They proceed on the expectation that the weld was fine, because it usually is, and the programme cannot afford the wait. Most of the time that is correct. The times it is not are the times that end up in a failure investigation.
What we changed
The technician records results at the component, on a tablet, into a structured form built for the method being used. Not a photograph of a paper sheet — the actual fields, with the acceptance criteria for that job already loaded.
Several things follow from that.
The transcription step disappears entirely, and with it its errors. The data is entered once, by the person who took the reading, standing next to the thing it came from.
Validation happens at the point of entry. A reading outside the expected range prompts while the technician is still at the component and can check it, rather than three days later when they cannot.
Images attach where they were taken. Radiographs, scan captures, photographs of surface condition — bound to the specific test, not emailed separately and reunited later by someone guessing.
Review happens in parallel rather than in sequence. A Level 3 can open results as they are recorded, from anywhere, and query something while the crew is still on site. That single change removes the biggest queue in the old process.
And the client sees it. Approved results appear in the client portal as they are signed off, rather than arriving as a batch of PDFs days later.
What it looks like on a job
A recurring shape on fabrication work: a set of welds shot in the morning, interpreted through the day, results appearing against each weld as they clear. By the time the crew packs up, the fabricator knows which welds are released and which need attention — and the repair on the one that failed starts that afternoon instead of the following week.
The value is not that the report is faster. It is that the decision is faster, and the decision is what the programme was waiting for.
The obvious objections
"Faster reporting means less careful reporting." The opposite, in our experience. The review step is unchanged — the same Level 2 or Level 3 signs off against the same acceptance criteria. What changed is the waiting between steps, and removing a queue does not remove rigour. Entering data at the component is more careful than transcribing it later, not less.
"Our site has no signal." Plenty do not, particularly inside plant. The system works offline and synchronises when the device reconnects. The technician does not have to think about it.
"We need a formal PDF for our records." You still get one, generated from the same data, signed by the same person. The difference is that it is generated from the record rather than being the only record.
"Our people are not tablet people." The forms are built around the method, so the technician is answering the questions they would have answered on paper, in the same order. The training is short. The resistance is usually about the first week, and it does not survive contact with not having to do paperwork in the evening.
What it means for audit
A quieter benefit that matters on regulated work. Every result carries who recorded it, when, with which equipment, calibrated when, against which procedure and which acceptance criteria — because those fields are part of the record rather than a header typed onto a template.
When an auditor asks whether the technician who examined a particular weld held current certification in that method and sector on that date, the answer is a lookup rather than an archaeology project. On nuclear and pressure equipment work, where records outlive the people who made them, that is worth a great deal.
The honest limits
This does not make interpretation instant. A complex radiographic set still takes as long to interpret properly as it always did, and anyone promising otherwise is promising to look at it less carefully. What disappears is the queue around the interpretation, not the interpretation.
It also does not help if the acceptance criteria have not been agreed. Fast delivery of a result nobody can act on because the standard is still being argued about is not an improvement. The criteria go into the job before the crew mobilises — that part is process, not software.
Worth asking your current provider
Whoever does your inspection, three questions are worth putting to them:
How long between the technician finishing and us being able to act on the result?
How many times is that result transcribed between the gauge and the report?
If we need to know in five years who examined a specific weld and what they were certified in, how long does that take to answer?
If the answers are "three to five days", "twice", and "we would have to dig", there is time and money sitting in that gap.
We built the reporting we wanted to receive. If you would like to see it working on a real job rather than described, get in touch and we will show you.
not add more.
