Skip to main content

How to use Assisted import

Assisted Import is an AI-powered data import feature created to make importing information into Katana easier.

Written by David Lorbiecke

Assisted import is an AI-powered data import tool that allows you to quickly add items (products or materials), suppliers, customers, stock levels, and BOMs (BOMs are create only, no updating) by uploading your own spreadsheets (CSV, XLSX, or XLS) without having to use a pre-defined template.

With the import, you can also update products, materials, suppliers, customers, and stock (contacts and stock can be included in the same file as new items).

Each row is graded on upload, matched rows update existing records, while unmatched rows create new ones. Item matching works on Katana ID, then SKU, or name for items. For contacts, it's email and then name. BOMs are create-only.

The AI will automatically map your spreadsheet's column headers to Katana fields, validate the data, and flag potential errors so you can fix them before import.


Step-by-step guide

1. Open up Assisted import

  1. Navigate to Settings and find Data import from the left-side menu (under Import and more).

2. Upload your spreadsheet

  1. Choose what you'll be importing (not all options support Assisted import).

  2. Drag-and-drop the spreadsheet into the box or click Upload file

Note: If the uploaded spreadsheet contains multiple sheets, only the first one will be included.

3. Column mapping

  1. Each column in your file will be evaluated, and the AI will suggest mapping based on column names and data patterns.

    1. Each field mapped by the AI will also be tagged with a qualifier (High, Medium, Low) indicating the AI's confidence in its accuracy.

  2. Examine the mapping to verify that everything is correct. If fields are missing, drag-and-drop them from the unmapped column.

    Note: Required fields must be mapped before import can proceed.

  3. Once ready, click Proceed at the bottom of the screen.

4. Data validation

  1. This step validates each row in the file, ensuring it meets Katana’s rules (required fields, data types, value constraints, and uniqueness).

  2. Rows are split into two groups:

    1. Valid rows (green) which are ready to be imported

    2. Invalid rows (red) which need correcting or will be skipped.

      1. Only appears if there are invalid rows

      2. Cells can be updated by clicking on them

      3. A report can be downloaded from here to help show what is invalid

      4. Bulk fixes may be offered

Note: Assisted import matches each row to existing records and shows Create / Update / No changes / Flagged before commit.

If fields are being updated, these specific fields will be highlighted so that you know what is being updated before completing the import.

5. Import

  1. Click on the Import button

  2. The data is added to Katana.

    1. Processing happens in the background and completes quickly, even for larger files.

Note: If an item references a supplier, category, or unit of measure that doesn’t exist in Katana, it will be automatically created during import.


Limitations

Limitation

Details

Row limit: 5,000 per import

For larger files, we recommend splitting the file into batches or use the template-based import.

If the file is over the limit, the entire import will be rejected, there won't be a partial import.

Only the first sheet can be imported

If the spreadsheet contains multiple sheets, only the first sheet is read. We recommend moving all needed data to the first sheet or saving it as a separate file.

4 variant maximum

Assisted import doesn't support any products with more than 4 variant options. If there are more, you will need to use the template-based import.

Not supported Fields

Operations, batch and serial numbers, price lists, sales orders, custom fields, multi-bins, and BOM updates cannot be imported via the Assisted import. Use a template-based import instead.


Assisted import vs. Template-based import

Assisted import

Template-based import

Best for:
- Setting up Katana for the first time
- Existing spreadsheets that you don’t want to reformat

Best for:
- Updating BOMs

- Data Assisted import doesn't cover yet (operations, batch/serial stock, price lists, sales orders)

- large files (5,000+ rows).

Mapping: Automatic AI column mapping with confidence scores

Mapping: No automatic mapping, only fixed columns based on the template

Data validation: Can clearly see and fix errors before import

Data validation: If errors occur, an error report is created after an import attempt

Row limit: Up to 5,000

Row limit: Up to 50,000

Supported imports: Products, Materials, Suppliers, Customers, Stock levels, and BOMs (BOMs create-only)

Supported imports: All types except customers/suppliers and sales and purchase orders


Common issues during import

Behaviour

Why

If a row matches existing data in Katana, it is skipped

As nothing has changed, there's nothing to be updated. You will see "skipped to avoid duplication".

When running an update, if a row doesn't match an existing record, the row isn't rejected, new data is created instead.

On an update, anything that fails to match becomes a new record ("it added it as a new product")

Blank cells don't erase existing data

If a cell is empty, the system will just leave the data as it currently is

Unsure matches are flagged, not guessed

An SKU, name, or email matching two records, or two rows pointing at one record, gets set aside

A non-matched Katana ID skips the row

If the Katana ID doesn't match an existing one, the row is completely skipped

Only the first sheet is reviewed

Any other sheets will be completely ignored during import

Invalid rows are skipped, but the import still "succeeds"

A successful import means that it was completed, but not that everything was imported.

Comma-formatted numbers

1,000 and 1,5 only work for recipe quantities. Elsewhere they're invalid


Your feedback is invaluable. Let us know your thoughts on this article or anything in Katana you'd like to see improved: [email protected]

Did this answer your question?