Import candidates in bulk

The candidate import workspace supports resume batches and tabular data. Both paths include a review step so you can identify duplicates, invalid values, and incomplete records before creating candidates.

Import review is part of the workflow

Treat parsed and mapped information as a working draft. Confirm candidate identity, contact details, experience, and specialties against the source before relying on the imported record.

Prepare a batch you can review

Choose source records your agency is authorized to use and keep a recoverable original in the approved workspace. For a first import, use a small representative batch so the reviewer can inspect the result before larger volumes are introduced. Include the types of records the real workflow needs, without uploading unrelated personal information.

For spreadsheet data, use clear column headings and consistent formats. Inspect names, email addresses, phone numbers, and dates before mapping. A column called “location” may contain a candidate's residence, preferred market, or former employer's address; establish its meaning before selecting a destination field.

Treat mapping as a decision about meaning

Match each source column to the candidate information it actually represents. Leave unsupported or uncertain information out of the import until the responsible person resolves it. Do not place several different meanings into one field just because the destination accepts text.

Review the preview after mapping. Check examples from the beginning and later rows, especially if the file combines exports from different sources. A mapping that works for the first row may still misinterpret another part of the file. Correct the source or mapping before committing rows whose values would mislead recruiters.

Investigate duplicates and invalid rows

Compare possible duplicates with existing records using authorized identifying information. A duplicate warning needs review, not an automatic decision to create another person. For an invalid row, read the reported problem and correct the actual value. Exclude a row when its information cannot be established reliably.

Resume batches need the same care. Check that extraction associated employment, education, and contact details with the correct candidate. Low-confidence results remain review work even if the file uploaded successfully. Do not describe all uploaded resumes as completed candidate records while some still need attention.

Reconcile the result with the input

After committing, compare ready, created, skipped, and failed outcomes with the batch you reviewed. Open a sample of created records in Talent and compare them with the source. Confirm that the resulting fields support the intended recruiting task before starting matching or outreach.

If the response is interrupted, inspect the saved results before importing the same file again. Repeating the entire batch can make it harder to distinguish records already created from unresolved rows. Keep a private record of the import time, source version, mapping decisions, and rows needing follow-up. Use the skipped-row help article for recovery, and share source files with support only through the agreed secure process.