Parcha Testing Best Practices Guide

Last updated: July 22, 2025

Introduction

Welcome to the Parcha Testing Best Practices Guide! This guide will help you effectively test your compliance agents and ensure they're working correctly before deploying them in production. Parcha provides powerful AI-driven compliance agents that can perform various checks including AML screening, due diligence, and risk assessment.

What You'll Learn

How to quickly test agents using built-in sample data (no setup required!)
How to set up and run test cases with your own data
Best practices for different types of compliance agents
How to use CSV and JSON data formats effectively
Common troubleshooting techniques

Getting Started with Testing

Prerequisites

Before you begin testing, ensure you have:

Access to your Parcha account (you can sign up at parcha.ai)
API credentials (if testing via API)
Sample data in the correct format:
- CSV format for web portal uploads
- JSON format for API testing

Testing Environment

Parcha provides a sandbox environment where you can safely test your agents without affecting production data. When you log in, you'll see the Agent Hub where you can manage and test your compliance agents.

Quick Start Guide

Fastest way to test (< 2 minutes):

Log in to Parcha: Navigate to your Parcha dashboard
Access Agent Hub: Click on "Agent Hub" in the sidebar
Select an Agent: Choose any agent you want to test
Click "Run Cases" → "Sample Data Set"
Pick a scenario and click "Run"

That's it! Your test case will start processing immediately with realistic data.

Understanding the Agent Hub Interface

When you access the Agent Hub, you'll see:

Template Section: Pre-built agent templates for quick setup
- Business AML Screening
- Enhanced Business Due Diligence
- Individuals AML Screening
- MCC Code Generation
Your Compliance Agents: Custom agents you've created
- Shows agent status (Draft/Active)
- Case count and last updated information
- Quick access to agent configuration
Pre-configured Agents: Organization-specific agents
- Ready-to-use agents for specific compliance needs
- Configured according to your organization's requirements

Understanding Agent Types

Parcha offers several types of compliance agents, each designed for specific use cases:

1. Business AML Screening

Purpose: Performs anti-money laundering checks on businesses
Key Features:
- Sanctions screening
- Adverse media checks
- Risk rating assessment

2. Enhanced Business Due Diligence

Purpose: Comprehensive investigation of businesses
Key Features:
- Online presence verification
- Address validation
- Business owner identification
- High-risk industry detection

3. Individuals AML Screening

Purpose: Screens individuals for compliance risks
Key Features:
- PEP (Politically Exposed Person) screening
- Sanctions checks
- Adverse media monitoring
- Career history verification

4. MCC Code Generation

Purpose: Generates and verifies Merchant Category Codes
Key Features:
- Automatic MCC code assignment
- Risk rating based on business type
- Verification against self-attested data

Using Sample Data

Quick Start with Built-in Sample Data

The easiest way to test your agents is using Parcha's built-in sample data:

Navigate to your agent in the Agent Hub
Click "Run Cases" or "Add Cases" dropdown
Select "Sample Data Set" option
Choose from pre-configured test scenarios

This instantly loads realistic test data tailored to your agent type - no manual data creation needed!

Available Sample Data Sets

Parcha provides different sample data scenarios for comprehensive testing:

For Business (KYB) Agents:

Standard Business: Clean business with typical profile
High-Risk Business: Businesses in regulated industries
International Business: Non-US entities with cross-border operations
Complex Structure: Multiple owners, subsidiaries, and UBOs

For Individual (KYC) Agents:

Standard Individual: Clean profile with no adverse findings
PEP Individual: Politically exposed persons
High-Risk Individual: Individuals with potential compliance concerns
Common Name: Test false positive handling

Each sample data set includes:

Complete profile information
Realistic addresses and identifiers
Associated individuals/entities (where applicable)
Document references
Various risk profiles

Custom Sample Data Format

If you need to test with your own data, Parcha accepts different formats:

Web Portal: CSV format
API: JSON format

Let's look at examples for both formats:

CSV Format (Web Portal)

KYC (Individual) CSV Template

first_name,middle_name,last_name,name_prefix,name_suffix,date_of_birth,address.street_1,address.street_2,address.city,address.state,address.postal_code,address.country,associated_addresses.0.street_1,associated_addresses.0.street_2,associated_addresses.0.city,associated_addresses.0.state,associated_addresses.0.postal_code,associated_addresses.0.country,country_of_nationality,country_of_residence,place_of_birth,gender,email,phone,title,linkedin_profile_url,current_employer,employer_industry,current_job_title,is_applicant,is_business_owner,proof_of_address_documents.0.document_type,proof_of_address_documents.0.document_url,business_ownership_percentage,source_of_funds_description,source_of_funds_documents.0.document_type,source_of_funds_documents.0.document_url,source_of_funds_amount
John,M,Doe,,,1990-01-15,123 Main St,,San Francisco,CA,94103,US,,,San Francisco,CA,94103,US,US,US,"San Francisco, USA",Male,john.doe@example.com,5551234567,Software Engineer,,,Software,TRUE,FALSE,utility_bill,https://example.com/document.pdf,0.0,,,
Jane,,Smith,,,1985-05-20,456 Market St,,New York,NY,10001,US,,,New York,NY,10001,US,US,US,"New York, USA",Female,jane.smith@example.com,5559876543,Product Manager,,,Technology,TRUE,TRUE,utility_bill,https://example.com/document.pdf,0.3,,,

KYB (Business) CSV Template

business_name,registered_business_name,address_of_operation,address_of_incorporation,website,business_purpose,description,industry,tin_number,partners,customers,source_of_funds,customer_countries,incorporation_date,business_registration_number,contact_email_address,contact_phone_number,associated_individuals.0.first_name,associated_individuals.0.middle_name,associated_individuals.0.last_name,associated_individuals.0.date_of_birth,associated_individuals.0.email,associated_individuals.0.phone,associated_individuals.0.country_of_nationality,associated_individuals.0.country_of_residence,associated_individuals.0.title,associated_individuals.0.is_business_owner,associated_individuals.0.business_ownership_percentage,associated_individuals.0.address.street_1,associated_individuals.0.address.street_2,associated_individuals.0.address.city,associated_individuals.0.address.state,associated_individuals.0.address.postal_code,associated_individuals.0.address.country_code
Acme Corporation,Acme Corp LLC,"123 Main St, San Francisco, CA 94103","456 Corporate Blvd, Wilmington, DE 19801",www.acmecorp.com,Technology Solutions,Enterprise software development company,Software,12-3456789,Microsoft,Fortune 500 companies,Series B funding,US,2015-03-20,C123456,info@acmecorp.com,+1-415-555-0100,John,,Smith,1975-08-15,john.smith@acmecorp.com,+1-415-555-0101,US,US,CEO,true,60,123 Founder Way,,San Francisco,CA,94103,US
Global Logistics LLC,Global Logistics Limited,"456 Market St, London, UK","456 Market St, London, UK",www.globallogistics.com,Freight Shipping,International freight and logistics services,Logistics,987-654-3210,DHL,Manufacturers,Investments,United Kingdom,2019-05-15,BRN-234567,info@globallogistics.com,+44-20-1234-5678,David,R,Brown,1980-01-20,david.brown@globallogistics.com,+44-20-1234-5678,UK,UK,Managing Director,true,40,10 Downing Street,,London,,SW1A 2AA,UK

Important Notes about CSV Format:

Headers must match exactly - The system expects specific column names as shown above
Dates - Use YYYY-MM-DD format (e.g., 1990-01-15)
Boolean values - Use TRUE/FALSE (case insensitive)
Empty values - Leave blank or use empty quotes
Addresses - Can be provided as:
- Single field: "123 Main St, San Francisco, CA 94103" (will be auto-parsed)
- Separate fields: address.street_1, address.city, etc.
Associated individuals - Use indexed notation: associated_individuals.0.first_name, associated_individuals.1.first_name, etc.

Intelligent Column Mapping

Parcha uses AI to automatically map your CSV columns to the correct fields. Common variations are handled automatically:

Automatic Mappings:

fname → first_name
lname → last_name
DOB → date_of_birth
company → business_name
email_address → email (KYC) or contact_email_address (KYB)
phone_number → phone (KYC) or contact_phone_number (KYB)
linkedin → linkedin_profile_url
EIN → tin_number

The system will show you suggested mappings with confidence scores, allowing you to review and adjust before importing.

JSON Format (API)

Business Entity Sample

{"id": "test-business-001","self_attested_data": {
    "business_name": "Acme Corporation",
    "registered_business_name": "Acme Corp LLC",
    "website": "https://www.acmecorp.com",
    "ein_number": "12-3456789",
    "industry": "Technology",
    "address_of_operation": {
      "street_1": "123 Main Street",
      "street_2": "Suite 100",
      "city": "San Francisco",
      "state": "CA",
      "country_code": "US",
      "postal_code": "94105"
    },
    "address_of_incorporation": {
      "street_1": "456 Corporate Blvd",
      "city": "Wilmington",
      "state": "DE",
      "country_code": "US",
      "postal_code": "19801"
    }}}

Individual Entity Sample

{"id": "test-individual-001","self_attested_data": {
    "first_name": "John",
    "middle_name": "Michael",
    "last_name": "Smith",
    "date_of_birth": "1985-03-15",
    "email": "john.smith@example.com",
    "phone": "+14155551234",
    "address": {
      "street_1": "789 Oak Avenue",
      "city": "Los Angeles",
      "state": "CA",
      "country_code": "US",
      "postal_code": "90001"
    },
    "country_of_nationality": "US",
    "country_of_residence": "US",
    "is_business_owner": true,
    "business_ownership_percentage": 51}}

Business with Associated Individuals

{"id": "test-business-002","self_attested_data": {
    "business_name": "Global Trading Inc",
    "website": "https://globaltrading.com"},"associated_individuals": [
    {
      "id": "owner-001",
      "self_attested_data": {
        "first_name": "Sarah",
        "last_name": "Johnson",
        "title": "CEO",
        "is_business_owner": true,
        "business_ownership_percentage": 60
      }
    },
    {
      "id": "owner-002",
      "self_attested_data": {
        "first_name": "Michael",
        "last_name": "Chen",
        "title": "CTO",
        "is_business_owner": true,
        "business_ownership_percentage": 40
      }
    }]}

Real-World Test Example

Here's a complete test case example based on Parcha's demo data:

{"id": "parcha-labs-test","self_attested_data": {
    "business_name": "Parcha Labs, Inc.",
    "registered_business_name": "Parcha Labs, Inc.",
    "website": "https://parcha.ai",
    "ein_number": "12-3245321",
    "industry": "Artificial Intelligence, FinTech",
    "address_of_operation": {
      "street_1": "1160 Battery St Suite 100 #1014",
      "city": "San Francisco",
      "state": "CA",
      "country_code": "US",
      "postal_code": "94111"
    }},"associated_individuals": [
    {
      "id": "founder-1",
      "self_attested_data": {
        "first_name": "John",
        "last_name": "Doe",
        "title": "CEO",
        "is_business_owner": true,
        "business_ownership_percentage": 50
      }
    }]}

Running Test Cases

Via Web Interface

Navigate to Agent Hub
- Click on "Agent Hub" in the sidebar
- You'll see all available agents organized in three sections:
  - Create new AI Agent templates
  - Your compliance agents
  - Pre-configured compliance agents
Select an Agent
- For new agents: Click on a template card (e.g., "Business AML Screening")
- For existing agents: Click on the agent name or case count
- Active agents show a green "Active" badge
- Draft agents show an orange "Draft" badge
Create a Test Case
- Click "Run Cases" or use the "Add Cases" dropdown
- Choose your data input method:
  - Sample Data Set (Recommended for testing): Pre-configured test scenarios
  - Upload CSV: Drag and drop or browse for your CSV file
  - Manual Entry: Fill out the structured form
  - Batch Upload: Upload multiple cases via CSV for bulk testing
Configure Test Parameters
- Set any agent-specific parameters
- Choose notification preferences
- Add case notes or tags for tracking
Monitor Progress
- Real-time status updates show each step:
  - Data validation
  - External data gathering
  - Analysis and risk assessment
  - Report generation
- Progress bar indicates completion percentage
- Any errors or warnings appear immediately
Review Results
- Summary View: Quick overview with risk scores
- Detailed Report: Full findings with evidence
- Data Sources: View all checked databases
- Download Options: PDF, JSON, or CSV formats

Via API

# Example API call to run a test case
curl -X POST https://api.parcha.ai/v1/agents/{agent_id}/cases \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "id": "test-case-001",
    "self_attested_data": {
      "business_name": "Test Company",
      "website": "https://testcompany.com"
    }
  }'

API Testing

Authentication

# Set your API key as an environment variableexport PARCHA_API_KEY="your_api_key_here"

Running Batch Tests

import requests
import json

# Load test datawith open('test_data.json', 'r') as f:
    test_cases = json.load(f)

# Run testsfor case in test_cases:
    response = requests.post(
        f"https://api.parcha.ai/v1/agents/{agent_id}/cases",
        headers={
            "Authorization": f"Bearer {api_key}",
            "Content-Type": "application/json"
        },
        json=case
    )
    print(f"Test case {case['id']}: {response.status_code}")

Checking Test Results

# Get case status
curl -X GET https://api.parcha.ai/v1/cases/{case_id} \
  -H "Authorization: Bearer YOUR_API_KEY"

Best Practices by Agent Type

Business AML Screening

Do's:

✅ Provide complete business information including legal name and registration details
✅ Include website URL for online presence verification
✅ Test with businesses from different industries and jurisdictions
✅ Include edge cases (newly formed businesses, foreign entities)

Don'ts:

❌ Use incomplete or abbreviated business names
❌ Skip address information
❌ Test only with US-based entities

Test Scenarios:

Standard Business: Regular corporation with clean history
High-Risk Industry: Cryptocurrency exchange, money services
Foreign Entity: Non-US business with limited information
Shell Company: Minimal online presence, recent incorporation

Enhanced Business Due Diligence

Do's:

✅ Include all available business identifiers (EIN, registration numbers)
✅ Provide both operational and incorporation addresses
✅ Include associated individuals with ownership percentages
✅ Test with complex corporate structures

Don'ts:

❌ Omit beneficial ownership information
❌ Use outdated business information
❌ Skip website verification for online businesses

Test Scenarios:

Simple Structure: Single owner, single location
Complex Structure: Multiple owners, subsidiaries, parent companies
International Business: Operations in multiple countries
High-Risk Profile: Business in sanctioned country or high-risk industry

Individuals AML Screening

Do's:

✅ Use complete legal names (first, middle, last)
✅ Include date of birth for accurate matching
✅ Provide current address information
✅ Test with common names to check false positive handling

Don'ts:

❌ Use nicknames or abbreviated names
❌ Skip nationality/residence information
❌ Test only with unique names

Test Scenarios:

Clean Individual: No adverse findings expected
PEP: Politically exposed person or family member
Common Name: John Smith, Maria Garcia (high false positive potential)
Sanctioned Individual: Person on sanctions lists

MCC Code Generation

Do's:

✅ Provide detailed business description
✅ Include product/service information
✅ Test with ambiguous business types
✅ Verify against self-attested MCC codes

Don'ts:

❌ Use vague business descriptions
❌ Skip industry classification
❌ Ignore risk rating validation

Troubleshooting Common Issues

Issue: "Invalid Data Format"

Solution:

Verify JSON syntax is correct
Check all required fields are present
Ensure date formats are YYYY-MM-DD
Validate country codes are ISO 3166-1 alpha-2

Issue: "Agent Not Responding"

Solution:

Check agent status is "Active"
Verify API credentials are valid
Ensure you're not exceeding rate limits
Check system status at status.parcha.ai

Issue: "Incomplete Results"

Solution:

Verify all required data fields are populated
Check for data quality issues (typos, formatting)
Ensure external data sources are accessible
Review agent configuration settings

Issue: "High False Positive Rate"

Solution:

Provide more specific identifying information
Include middle names and DOB for individuals
Use full legal business names
Add additional context (website, address)

Advanced Testing Strategies

1. Regression Testing

Create a suite of standard test cases that you run whenever:

Agent configuration changes
New features are added
System updates occur

2. Edge Case Testing

Test with:

Minimal required data
Maximum data fields
Special characters in names
Non-Latin scripts
Unusual business structures

3. Performance Testing

Test with batch uploads
Monitor processing times
Check rate limits
Validate concurrent case handling

4. Integration Testing

If using Parcha with other systems:

Test webhook notifications
Validate API response formats
Check data synchronization
Verify error handling

5. Compliance Testing

Ensure your testing covers:

Different risk levels
Various jurisdictions
Regulatory requirements
Documentation standards

Testing Checklist

Before going to production, ensure you've tested:

Initial Testing (Using Sample Data Sets):

[ ] Run each agent with at least one sample data scenario
[ ] Test different risk profiles using sample data (Standard, High-Risk, PEP, etc.)
[ ] Verify report generation with sample data
[ ] Review how your agent handles various sample scenarios

Advanced Testing (With Your Data):

[ ] All agent types you plan to use
[ ] Various risk profiles (low, medium, high)
[ ] Different geographic regions
[ ] Edge cases and error scenarios
[ ] API integration (if applicable)
[ ] Batch processing capabilities
[ ] Report generation and export
[ ] Webhook notifications (if configured)
[ ] User access and permissions
[ ] Data retention and privacy compliance

Resources and Support

Documentation

Support Channels

Email: support@parcha.ai
Documentation: https://docs.parcha.ai
Status Page: https://status.parcha.ai

Community

Join our community forum for tips and best practices
Share your testing strategies with other users
Get updates on new features and improvements

Conclusion

Effective testing is crucial for ensuring your compliance processes work smoothly. By following these best practices, you can:

Reduce false positives
Improve accuracy
Ensure regulatory compliance
Optimize processing times

Remember to test thoroughly with representative data before deploying any agent to production. Happy testing!

Last updated: January 2025 Version: 1.0