Every fixed-asset file names its columns differently. One calls it "Date in Service", the next "PIS Date", a third "Acquired". Mapping is the step where you say which of your columns means what.
It's mostly done before you see it. This article covers what the step is for and what it can and can't do to your import.
Where this is in the app
Dashboard → Import assets — or, from a return, Assets tab → Add assets → Bulk import
Mapping is the Map Columns step of the import: step 2 with your own file, step 3.5 on the vendor route. The two doors are compared below.
Your columns, our fields
Whatever you upload becomes the same thing by the time you see it: a table of columns and rows. Your own spreadsheet, a vendor's export, a PDF schedule. The mapping step is the same table on every route in, so learning it once covers all of them.
Three fields are required. Description, placed-in-service date, and original cost.
Everything else is optional. Tax, book, AMT/ACE, vehicle, organization and disposal fields. Map what your file happens to carry.
Anything unmapped is ignored. A column you don't map isn't imported. That's a legitimate answer, not a gap.
Mapping cannot block your import
No mapping decision refuses to let an import proceed.
Columns you leave unmapped are left behind, and that's a normal outcome. Vendor files routinely carry columns no fixed-asset system has a field for.
What stops an import is a row missing data it actually needs, and that's a later step, judged on the rows themselves.
Two doors, one table
There are two ways into an import and they meet at the same table with the same controls. What differs is how much has been decided before you get there.
Your own file | A vendor export | |
Where | Import, step 2 | Third-party import, step 3.5 |
You see it | Always | Unless the mapping is complete and the AI agrees with the vendor |
Starting point | Synonym matching, then AI suggestions | The vendor's known layout, pre-applied |
Badges | HIGH, Suggested | HIGH, MEDIUM, LOW, plus Use AI mapping |
Editing | The same table, dropdowns and views | The same table, dropdowns and views |
The vendor route is quieter, not silent. When a file's layout is already known, asking you to confirm sixty columns you didn't choose isn't carefulness, it's a toll.
What it does still stop for is disagreement. Your file is parsed twice, once by the vendor's adapter and once by our AI, and where the two read a column differently the step appears so you can pick. The column says which is which, for example "Fixed Assets CS (Thomson Reuters) adapter mapped this column to OriginalMethodCode; AI suggests Method", with a Use AI mapping button. Do nothing and the vendor's reading stands.
This is common enough that a clean, fully recognized export still often stops here.
Getting back to it
On the vendor route, the preview screen has a Back to mapping control. Reach the preview, notice a column landed somewhere odd, and you can go straight back with your edits intact. That works even on a file that went straight through without stopping at step 3.5 at all.
Both routes produce the same kind of object and hand it to the same table, which is why a fix made to one shows up on the other.