Overview
An observation is a site-safety record: something seen on a site that deserves attention — a trip hazard, a missing barrier, an unsafe practice — captured with photos and worked through a review lifecycle until it's dealt with. Where a defect is anchored to a machine, an observation is anchored to a site; between them, "I saw something wrong" always has a home.
Before you start
- Plan
- Builder — observations belong to site-hosting plans
- Role
- Site personnel (officers and managers) record and work them via the Go app; console roles manage them from the office
- A site — every observation belongs to one
Recording an observation
From the Go app (the site team's web app) or the console: pick the site, describe what you saw, attach photos, and submit. The observation lands in Reported, attributed and timestamped, and the site's responsible people are notified.
The review lifecycle
REPORTED → REVIEWED → ACTIONED → CLOSED
↑
DISMISSED (from Reported or Reviewed; re-reviewable)
| Status | Meaning |
|---|---|
| Reported | Captured, awaiting review |
| Reviewed | Assessed by someone with authority — real, and being handled |
| Actioned | The corrective action has been taken |
| Closed | Signed off. Terminal |
| Dismissed | Reviewed and judged not to require action — reversible: a dismissed observation can be brought back to Reviewed |
Two properties worth noting: dismissal is honest and recoverable — it records "we looked and decided no action", and it can be reopened if that judgement changes; and closure is terminal — a recurrence is a new observation.
Every transition is permission-checked against both console roles and site roles — a site manager can work their site's observations without needing office authority, and vice versa — and every transition is recorded immutably.
Assignment, comments, and escalation
An observation can be assigned to the person who'll fix it (they're notified), discussed in its comment thread, and escalated when it needs more senior attention — escalation notifies up the chain rather than quietly re-sorting a list.
Observation, defect, or spot check?
| You have… | Record a… |
|---|---|
| A fault on a machine | Defect — it gates the machine and drives repair |
| A site condition or practice needing attention | Observation — site-anchored, review lifecycle |
| An oversight record — "I checked, here's the outcome" | Spot check — append-only, no lifecycle |
The boundaries are worth keeping clean: a machine fault recorded as an observation never gates the machine — only a defect does that.
Edge cases
- Contractors on your site — observations involving a contractor's people or plant are still your site's observations; the site's review chain owns them, whatever's being observed.
- Photo evidence — attach at capture and again at actioned, so the record shows before and after.
- Where they show — observations are listed on the console's Site Observations page and in the site team's queue in the Go app. They don't appear on the dashboards or count towards a site's compliance figures.
Troubleshooting
| Symptom | Likely cause | What to do |
|---|---|---|
| No observations feature visible | Not a site-hosting plan, or no site access role | Observations are a Builder capability worked by site personnel |
| Can't transition an observation | The move needs a higher console or site role | Check the site role — reviewing and closing sit above reporting |
| Dismissed something that turned out real | Dismissal is reversible | Bring it back to Reviewed and work it |
| A machine fault sitting as an observation | Wrong record type — observations don't gate machines | Raise a defect on the machine, reference the observation |