> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dibby.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Human Validation

> Add human oversight to your automated workflows

## Overview

<div align="center">
  <img src="https://mintcdn.com/dibby/pqX3-_kQLdN967hk/images/human-validation-screenshot.webp?fit=max&auto=format&n=pqX3-_kQLdN967hk&q=85&s=144787393fff2875782f7bbd1d7f3017" alt="Human validation" width="3456" height="1988" data-path="images/human-validation-screenshot.webp" />
</div>

Human validation allows you to pause workflow execution for manual review and approval. This is essential for workflows that require human judgment or verification.

## When to use human validation

Use human validation when:

* Processing sensitive or high-value documents
* Making critical business decisions
* Verifying AI extraction accuracy
* Compliance or regulatory requirements demand human oversight
* Data quality needs manual verification

## Validation types

### Optional validation

AI nodes like **Document Extractor** support optional validation:

* Workflow continues automatically by default
* Can be flagged for human review if needed
* Useful for quality checks on a subset of runs

### Mandatory validation

The **Manual Validation** node always requires human input:

* Workflow pauses at this node
* Workflow resumes after approval

## Validation process

1. **Workflow Pauses**: Execution stops at the validation node
2. **Task Created**: A validation task is created
3. **Review Data**: Reviewer examines data and reference documents
4. **Make Corrections**: Reviewer can edit extracted data
5. **Approve**: Reviewer approves and workflow continues

## Validation interface

The validation interface shows:

* Original document (if applicable)
* Extracted/processed data
* Expected values for comparison
* Edit capabilities for corrections
* Approve/reject buttons

## Best practices

* Assign validation tasks to specific team members
* Set up notifications for pending validations
* Monitor validation queue regularly
* Document validation guidelines for consistency
* Use optional validation for lower-risk items
* Reserve mandatory validation for critical data
