---
sourceDocument: Brazil Enable AI
sourceDocumentLink: https://www.servicenow.com/docs/r/intelligent-experiences

 Release :

    - brazil

ft:locale :

    - en-US

ft:publication_title :

    - Brazil Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# Extraction methods

# Content extraction methods {#ariaid-title1}

Release version: Brazil  
Updated November 5, 2026  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 2 minutes to read  
The Extract information from documents skill (Information Extraction skill) supports three extraction methods---field, table, and Q\&A---each suited for different types of document data. Understanding which method to use helps
you capture and store the right information in ServiceNow.

The Information Extraction skill uses AI-assisted parsing to extract data from documents (invoices, purchase orders, and contracts) and then populates records in ServiceNow. Select a method or
combination of methods that matches the type of information you want to capture and store in ServiceNow tables.{#content-extraction-methods__data-extraction-methods-para-1}

## Field extraction {#content-extraction-methods__data-extraction-methods-field}

Field extraction captures specific, isolated data points from a document---values that appear exactly once---and maps each to a single field on a ServiceNow record.{#content-extraction-methods__data-extraction-methods-field-para-1}

Use field extraction for data such as:{#content-extraction-methods__data-extraction-methods-field-para-2}

* Invoice numbers, purchase order numbers, or document identifiers
* Vendor or customer names
* Dates, such as an invoice date, due date, or contract start date
* Monetary totals or tax amounts
* Addresses or other contact details

{#content-extraction-methods__ul_ztx_z31_ckc}

You can bundle related single fields into a logical group. For example, a Billing Address group can capture Street, City, State, and ZIP from the same region of a document.{#content-extraction-methods__data-extraction-methods-field-para-3}

## Table extraction {#content-extraction-methods__data-extraction-methods-table}

Table extraction captures dynamic, multi-row data from documents where the number of entries varies between documents, such as invoice line items or expense report details. A table acts as a
two-dimensional grid of rows and columns that captures a list of related records.{#content-extraction-methods__data-extraction-methods-table-para-1}

Use table extraction for data such as:{#content-extraction-methods__data-extraction-methods-table-para-2}

* Invoice or purchase order line items (quantity, description, unit price, total)
* Expense report breakdowns
* Parts lists or bills of materials
* Multi-page flight itineraries or timesheet entries
{#content-extraction-methods__ul_a5x_z31_ckc}

## Q\&A extraction {#content-extraction-methods__data-extraction-methods-qa}

Q\&A extraction uses generative AI to answer open-ended, natural language questions about the contents of a document. It is suited for unstructured text that does not fit neatly into a
predefined field or table schema. The extraction processes meaning rather than matching a fixed pattern.{#content-extraction-methods__data-extraction-methods-qa-para-1}  
Warning:  
Q\&A extraction uses generative AI and may produce inaccurate or incomplete results. Review all AI-generated output before using it to populate records.

Use Q\&A extraction for questions such as:{#content-extraction-methods__data-extraction-methods-qa-para-2}

* The reason for termination
* Special shipping or delivery terms
* The designated project manager
* Narrative clauses or conditional language from contracts

{#content-extraction-methods__ul_b5x_z31_ckc}

You can also use Q\&A extraction to categorize documents.{#content-extraction-methods__data-extraction-methods-qa-para-3}

## Selecting an extraction method {#content-extraction-methods__data-extraction-methods-select}

Use the following guidance to choose the appropriate extraction method.{#content-extraction-methods__data-extraction-methods-select-para-1}
{#content-extraction-methods__table_tx1_p3v_wjc__entry__3}

| Extraction method | When to use | Example |
|-|-|-|
| Field extraction | Use when the data point appears once and has a well-defined type. | A number, date, or name. |
| Table extraction | Use when the document contains a list or grid of repeating items where the row count is unknown or varies across documents. | Line items in an invoice or purchase order. |
| Q\&A extraction | Use when the information is embedded in narrative text and requires interpretation rather than pattern matching. | An answer based on contract terms or policy text. |
[Table 1. Extraction method selection guidance]

{#content-extraction-methods__table_tx1_p3v_wjc}

You can combine multiple methods within a single use case to handle all of the data types present in a document.{#content-extraction-methods__data-extraction-methods-select-para-2}
**Related concepts**   

* [Use cases for the Information Extraction skill](https://www.servicenow.com/docs/ypWz7j8CDtyvZwajvBX1yQ "Use cases in the Information Extraction skill define what documents to process and what information to extract, so generative AI can accurately retrieve structured data from unstructured documents.")