A Step-by-Step Guide to creating your first AI agent for compliance in minutes

Last updated: July 17, 2025

Agent Hub allows compliance teams to create, customize, and deploy AI agents to automate manual workflows in minutes. This guide walks you through the complete process of building your first AI agent using our Enhanced Due Diligence template.

Overview

AI agents can transform time-consuming compliance processes from manual, multi-day workflows into automated, hour-long operations. This tutorial demonstrates how to build an Enhanced Due Diligence (EDD) agent that can conduct business research, screen for adverse media, verify addresses, and provide risk ratings automatically.

Getting Started

First go to hub.parcha.ai and sign-up in order to setup you Agent Hub workspace then follow the steps below...

Step 1: Select Your Template

Begin by choosing a pre-built template that matches your workflow requirements:

  1. Navigate to the template library in Agent Hub

  2. Browse available templates for your specific use case

  3. Select the template that best fits your needs (for this example, we'll use "Enhanced Due Diligence")

Templates provide a foundation with pre-configured steps that you can customize rather than building from scratch.

Step 2: Configure Basic Agent Information

Set up your agent's identity and purpose:

  1. Agent Name: Provide a descriptive name (e.g., "Fast Bank EDD Agent")

  2. Description: Add context about your business and the agent's role

  3. Purpose: Clearly define what the agent should accomplish

This information helps the AI understand your business context and make more informed decisions.

Configuring Agent Steps

Step 3: Set Up Core Workflow Steps

The Enhanced Due Diligence template includes these default steps:

  • Business Web Research: Gathers information about the target business

  • Business Profile Creation: Compiles a comprehensive business overview

  • Adverse Media Screening: Searches for negative news or reports

  • Address Verification: Confirms business location details

  • Industry Classification: Categorizes the business type

  • Country Classification: Identifies operational regions

  • Business Review Analysis: Evaluates customer feedback and ratings

  • Business Owner Verification: Confirms ownership information

Step 4: Customize Your Steps

Each step in the workflow can be customized.

For Business Web Research you can choose research depth:

  • Fast Research: Focuses on the business website only (faster results)

  • Deep Research: Examines the complete online presence including social media and reviews (more comprehensive)

For enhanced due diligence, deep research is recommended to gather sufficient information for subsequent analysis steps.

For Adverse media you can choose a number of options.

  1. Data Sources: Select from available providers (Google News, Google Search, specialized databases)

  2. Matching Criteria:

    • Choose exact match or partial match for names

    • Select geographic matching level (country, state, or city)

    • For business screening, country-level matching is typically most relevant

  3. Topics: Select specific adverse media categories to monitor

  4. Filtering Options: Configure settings to exclude results where the business is not the perpetrator

Step 5: Add Additional Verifications

Enhance your agent with supplementary checks:

  • Legal Entity Status Verification: Confirms business registration and standing

  • Sanction Screening: Checks against regulatory watchlists

  • Merchant Code Classification: Categorizes business type for payment processing

Add these components based on your specific compliance requirements.

Risk Assessment Configuration

Step 7: Create Risk Rating Criteria

Define how your agent should assess and categorize risk:

  1. High Risk Criteria: Define conditions that automatically flag businesses as high risk

  2. Medium Risk Criteria: Set parameters for cases requiring human review

  3. Low Risk Criteria: Establish conditions for automatic approval

Example criteria:

  • Any adverse media findings = High Risk

  • Incomplete verification = Medium Risk

  • Clean results with full verification = Low Risk

You can customize these criteria by typing new rules or modifying existing ones.

Deployment and Testing

Step 8: Review and Create Your Agent

Before deployment, review:

  • All configured steps and their parameters

  • Required input fields

  • Credit requirements for operation

  • Estimated processing time

The system will display the total credits needed to run your agent and provide guidance on subscription plans based on your expected usage volume.

Step 9: Test with Sample Data

Before processing real customer data:

  1. Select "Test Sample Data" from the options

  2. Choose sample data that matches your typical customer profile

  3. Select specific test cases from the provided library

  4. Review credit requirements for your test run

  5. Execute the test to validate your agent's performance

This approach allows you to refine your agent's configuration without using sensitive customer information.

Step 10: Monitor Real-Time Execution

Once your agent begins processing:

  • Monitor progress through the real-time dashboard

  • Observe each step as it completes

  • Review intermediate results and decisions

  • Verify the final risk assessment and supporting evidence

Alternative Input Methods

Agent Hub supports multiple ways to provide case data:

  • Sample Data: For testing and validation

  • CSV Upload: For batch processing multiple cases

  • Manual Entry: For individual case processing

  • API Integration: For automated workflow integration

Best Practices

Configuration Tips

  • Start with comprehensive research settings for initial testing

  • Gradually adjust parameters based on results and performance requirements

  • Regularly review and update risk criteria as your business needs evolve

  • Test with diverse sample data to ensure robust performance

Optimization Strategies

  • Monitor credit usage to optimize cost-effectiveness

  • Adjust research depth based on risk level requirements

  • Fine-tune adverse media criteria to reduce false positives

  • Regularly update screening parameters to reflect current regulatory requirements

Next Steps

Once your agent is operational:

  1. Monitor performance metrics and accuracy

  2. Gather feedback from compliance team members

  3. Refine risk criteria based on real-world results

  4. Consider expanding to additional use cases or workflows

  5. Integrate with existing systems via API connections

Support and Resources

For additional assistance:

  • Access our template library for other use cases

  • Review our API documentation for integration guidance: docs.parcha.ai

  • Contact support for customization requirements: support@parcha.ai

  • Join our community forum for best practices and tips

Your AI agent is now ready to transform your compliance workflows from manual processes into efficient, automated operations.