# Extract resume information

Use [Extract Information to Table](extraction-to-table.md) to organize selected resume fields into a table.

1. Prepare a small set of resumes.
2. Define fields such as name, education, and work experience, with descriptions.
3. Configure the result database, table name, and output location.
4. Run the workflow. Compare table fields with the originals, especially the structure of multiple education or employment entries.

The extracted structure must match the table. Describe how missing information should be handled and inspect those results.

For result files only, use [Document Information Extraction](document-extraction.md).

## Design the fields

Decide whether the target record includes multiple experience entries. Use a string for name and define education and employment as a single value or array. Use consistent dates and schema-compatible missing values.

Distinguish highest qualification from all education history in descriptions rather than using an ambiguous education field.

## Choose input scope

Table extraction combines inputs by default. A resume and its same-person attachments can contribute to one result. Do not combine several candidates into one record. Validate a single resume before configuring per-file batch processing and matching saving bindings.

## Inspect one result

| Content | Verification |
| --- | --- |
| Name and contact details | Compare with the original and check attachment attribution |
| Education and employment | Check counts, dates, and nested structure |
| Missing information | Retain agreed empty values without inventing details |
| Table | Check database, column types, and affected rows |
| Files | Locate this run's saved results |

## If writing fails

Check the extracted JSON against the schema, then inspect the table. Successful extraction does not prove a successful write, and table records do not prove later saving completed. Inspect existing data before changing columns or rerunning.

[Back to workflow templates](index.md)