AssetBase User Guide
DocsAB-024
Capturing Condition at Scan
Recording status and maintenance state when assets are scanned in the field
The moment you scan a tag is the only moment someone is guaranteed to be standing in front of the asset. That is the cheapest possible time to record what it actually looks like and update the fields only a person on site can know. This page covers what you can capture there, and the optional AI check that assesses the asset from a photo.
What You Can Capture
After a scan resolves to an asset, open Capture at scan (web: below the scan result; mobile: the Condition action on the scan confirmation, or the clipboard icon on any asset).
| Field | Notes |
|---|---|
| Condition | Good · Minor wear · Scratched · Dented · Damaged · Broken |
| Condition notes | What you can see — cracked housing, hydraulic leak |
| Found at | Where the asset actually is |
| Custom fields | Any of the asset's custom attributes, editable in place |
Press Save. AssetBase writes the asset and logs the scan as evidence in one step.
Condition is not maintenance state
They are deliberately separate:
- Condition is what the asset physically looks like right now.
- Maintenance state is where it sits in your workflow — operational, needs service, under repair, out of service.
An asset can be operational and scratched at the same time. Collapsing the two would lose exactly the information a stock-take exists to capture.
"Found at" versus expected location
Typing a Found at value records where the asset was actually seen. This is what drives the misplaced exception on a reconcile — the audit compares where you found it against where the register expected it.
It does not move the asset's expected home. If the asset genuinely lives there now, tick This is its new home. Without that, a stock-take could silently rewrite the very baseline it is auditing.
Capturing during an audit
Scanning into an open stock-take shows a Condition action on each capture confirmation. Taking it records the condition against that audit; ignoring it keeps the sweep at full speed. Condition captured this way is frozen into the audit's snapshot, so you can compare an asset's condition month over month from the stored reports.
Condition capture writes straight through and needs a connection. Offline, the scan itself is still queued and syncs later — you will be told the condition part needs signal rather than losing what you typed.
The AI Check
Press AI check and photograph the asset. AssetBase sends the photo — together with this asset's own register record — and gets back two things:
- Does it match? Whether the photographed item is plausibly the same kind of equipment as the record the tag resolved to, with a confidence and a one-line reason. This catches a tag on the wrong machine.
- What condition is it in? A grade from the same list you would pick by hand, with a confidence and the specific defects it can see.
The result fills the form in. Nothing is saved until you press Save. Review it, correct anything wrong, then save.
How to read the result
| What you see | What it means |
|---|---|
| Looks like this asset | The photo is consistent with the register record |
| Doesn't look like this asset | Worth investigating — a mis-applied tag, or a record that needs correcting |
| Inconclusive | The image was too dark, blurred or cropped to judge. No verdict is offered rather than a guess. |
Confidence is shown as a percentage. Treat a low-confidence verdict as a prompt to look more carefully yourself, not as an answer.
What gets recorded
Condition you accept from an AI assessment is stored marked ai_scan; condition you set yourself is stored as operator. A later reviewer can always tell a machine assessment from a human inspection.
Availability
The AI check needs internet and is cloud-only — an on-prem appliance is offline by design and will tell you so. It grades what is visible in the photograph; it does not open panels, test function, or read hour meters.