Writing a site visit report takes longer than the visit because by the time you sit down, you are no longer describing what you saw; you are reconstructing what it meant. The photograph survives the trip back to the office. The reason you took it does not. Three problems stack up between the site and the desk:
- Interpretation decays. Days later the photographs cannot explain why you took them, so you rebuild on Friday what you already knew on Tuesday.
- Notes get entered twice. Written on paper on site, typed into a computer at a desk: pure re-entry, at the moment the meaning has already faded.
- Drafting starts from a blank page. The template supplies the sections, not the reasoning, so the writing absorbs the framing you skipped, at a worse exchange rate.
The rest of this article takes them in order, and the fix is a sequencing change rather than a writing change.
The four steps a report goes through
Only one of the four is writing:
- Capture. You go on site and record the evidence: photographs, video, measurements.
- Framing. You attach meaning to the evidence while you still have it: why this photo, what the working theory is, what should be measured next.
- Drafting. The framed evidence becomes prose in the report's structure.
- Review and issue. The engineer reads, corrects, signs, and sends.
Most of the attention in this industry goes to step three, because that is the step that feels like work. Step two decides the cost. Framing has to happen while you are standing in front of the thing, or step three absorbs it later at a much worse exchange rate. The next section is why.
What decays between the site and the desk
What decays is not the memory of what you saw but the interpretation of what it meant. When you are standing in front of a wall, you know why you are photographing that particular spot. You have a working theory, a rough estimate, and the relevant standard near the front of your mind. None of that is in the photograph.
Now the usual path: you do not write it up that day, you attend one or two more site visits, and days later you sit down in front of a blank page, pull the pictures off your phone, and try to remember what you saw and, harder, what it meant. The photographs are all still there. The theory is gone, and rebuilding it takes longer than having it did.
Why framing gets skipped
Framing gets skipped because current tools make it feel like extra work rather than the same work moved earlier. Today's loop is: take photographs, write notes on paper, then transcribe those notes into a computer at a desk. The transcription step is pure re-entry, and it happens at exactly the moment the interpretation has already faded. Nobody comes to us asking to stop writing notes; they ask for a faster draft. The transcription is so normal that it has stopped registering as a step at all.
A better template does not fix this, either. A template tells you what sections the report has; it does not restore the reasoning you had on Tuesday and lost by Friday. The visit and the write-up are one job with a decay curve running between them, and every day of separation costs more than the day before.
Framing on site: narrate while you capture
Framing on site means narrating while you capture, so capture and framing become one act. Two paths, depending on what the site allows. Where you can shoot video, you talk while you film: the images get extracted from the video automatically, and your voice becomes the transcript. Where you cannot, and there are real reasons you often cannot, you take photographs and record a voice memo alongside them, and the notes write themselves. Either way there is nothing to transcribe, because there was never a written note to re-enter. Your interpretation was captured at the moment you had it, in your own words.
On the model's side, that narration is the substance. The model is instructed to work from your spoken interpretation rather than from the images, and it will not assert a defect you never called. Where that line sits, and what it means for a report going out under your seal, is worth reading in full. Framing on site moves when the interpretation is recorded, never who makes it: the engineer still decides what the defect is, what caused it, and what should be done, and the draft still gets read, corrected, and signed.
Time one write-up honestly this week
Measure one write-up in engineer-hours rather than elapsed days, and note how much of it was spent re-establishing what you already knew on site. Most engineers are surprised by the ratio. Then try one visit where you narrate as you capture: a voice memo alongside your photographs is enough, no new equipment needed. Write that one up the same day and compare. You are not testing whether talking to your phone feels natural. You are testing whether the second write-up needed reconstruction at all.
Where Tenera Reports removes the hours
Everything above works with a voice memo app and discipline, and that alone removes the reconstruction. Tenera Reports is drafting software built to collect the rest of the hours, and it maps onto the four steps directly:
- Capture and framing become one step. You photograph or film the site and talk while you do it, on the phone you already carry. There is no notebook to transcribe at a desk, because the transcript is made from your voice.
- Drafting stops being writing. The narration comes back as a draft in your firm's own template, with the relevant frames pulled from the footage and placed as figures. The office step changes from producing prose out of a blank page into editing a document that already says what you said, and editing is bounded in a way writing never is.
- Review gets a record. The reviewer reads the draft against the transcript of what was said on site, so checking becomes comparison rather than interrogation.
The steps that remain, walking the site, deciding what things mean, reading and signing, are engineering. They were never the hours that hurt.
Narrate one visit and compare
The first 15 reports in Tenera are free, with no time limit. Bring the report type you produce most often and one template, and narrate a single site visit rather than writing it up from memory. What you are checking is whether the draft that comes back needed reconstruction — everything else is detail.