Org Loom: Import CSV records onto the canvas

Before you begin

Each CSV needs a unique, nonblank header row. You will choose one Salesforce object for each file, then decide which columns contain Salesforce field values. Unmapped columns are skipped.

A row without a Salesforce ID becomes a draft, which is a new record that exists only on the canvas until you upload it. A row whose ID matches a record in the connected org can load that record and stage the CSV values as edits.

Connect the Salesforce org that contains any existing records.
Make sure the files are in CSV format and no larger than 50 MB each.
Give files distinct names so they are easy to identify.
Decide whether the import should add to the current canvas or replace it.
In a Team workspace, ask a workspace admin for the Upload to Salesforce permission if you will upload the staged rows.

This flow does not write to Salesforce. It stages the rows on the canvas. Use Upload to Salesforce later when you are ready. For very large jobs that do not need a canvas review, use Salesforce Data Loader.

The process

Open Import from CSV and add one or more files.
Choose an object and map Salesforce field columns under Columns.
Map Salesforce IDs or choose external-ID upsert when applicable.
Add the records to the current canvas or replace the canvas.
Review the resulting cards, then upload when ready.

Step 1: Add the CSV files

Open + Add records → Import from CSV. Drag files into the drop area or select the area to choose them from your device. You can load several files in one import.

Org Loom lists every accepted file separately. If a file is empty, unreadable, or not a CSV, the modal explains the problem and stays open so you can remove or replace that file.

Step 2: Choose the object and map columns

For each file, choose the Salesforce object represented by its rows. Org Loom may suggest an object from the filename, but you should confirm it before continuing.

Open Columns to review the mapping. The section starts closed so you can scan several files without a long list of fields. Headers that match a field label or API name may map automatically. A reference-style header ending in Key may also be suggested as a relationship column when Org Loom finds a confident match; always review the suggestion.

Each CSV column has three possible treatments:

Choose a Salesforce field when the CSV values should be stored in that field. A Salesforce field can receive values from only one column in the same file.
Choose Match to a related record in another CSV - not uploaded when the column is only a key for connecting records. You will finish that mapping under Relationships.
Choose Skip when the column should not be imported or used.

Lookup fields such as AccountId accept Salesforce IDs, not names or matching keys. To connect imported drafts, load the records and create their relationships on the canvas.

Map at least one Salesforce field in each file you want to import.
Salesforce-managed fields are not available as normal write targets.
If a mapped field is not writable for your Salesforce user, Org Loom warns you before the import and identifies values that will be left out.

Step 3: Choose how existing records are handled

To stage updates to records already in the connected org, map the CSV ID column to Salesforce ID (Id). Org Loom checks those IDs automatically; there is no separate “match and update” import mode.

A valid ID loads the current Salesforce record. CSV values are layered on top, and only values that differ are staged as edits.
A blank ID in the same file creates a new draft.
If an ID cannot be found in this org, Org Loom asks whether that row should become a new draft or be skipped. It is never silently inserted with the unresolved ID.

If the file maps an eligible Salesforce External ID field, the Operation control appears and offers Insert new records or Upsert by external id. Upsert is not a general CSV import choice: it is available only after an eligible External ID field is mapped. An External ID is a Salesforce field configured to identify a record using a value from another system. On upload, Salesforce updates a record with the matching value, or inserts a new record when there is no match. Review the selected field before using this option.

Connect records after importing

CSV import loads record values. It does not match rows by names, emails, or custom keys to create canvas links. Existing Salesforce lookup IDs remain field values; create connections between drafts on the canvas after loading them.

CSV export also contains record values only. It does not preserve links to draft records for re-import.

Connect records on a canvas

Step 4: Add or replace the canvas

Choose Add to canvas to keep the current records and place the imported records alongside them. Choose Replace canvas to start a new unsaved canvas from the import. Org Loom confirms before replacing existing work.

If a CSV ID matches a card already on the canvas, Org Loom does not create a duplicate. When the imported values differ, it opens Review imported changes so you can choose which values to copy onto the existing card. Closing that review without applying anything leaves the card unchanged.

Step 5: Review and upload

Inspect the imported cards before uploading. You can edit values, remove unwanted drafts, and add or correct relationships. Imported existing records show as modified only when the CSV staged a real change.

When the canvas is ready, choose Upload to Salesforce and review the upload preview and any validation messages. Drafts will be created, modified existing records will be updated, and unchanged existing records will be skipped.

Import safety and limits

The canvas holds up to 500 records. An import that would exceed that limit is refused before any records are added or replaced.
Incomplete or uncertain relationship mappings block the import; Org Loom does not silently choose a target or import an unresolved link.
After a successful import, the confirmation toast briefly offers Undo. Undo restores the records, relationships, and object selection from before the import.
If you edit affected records after the import, the older Undo is refused instead of overwriting your newer work.
Canceling before the final import leaves the canvas unchanged.