Vendor Data Integration
How to submit pre-existing vendor screening data to Parcha agents
Overview
When using certain Parcha agents (such as the IG AML Agent with ig-aml-vendor-v1-demo
), you can provide pre-existing screening data from compliance vendors like ComplyAdvantage or Refinitiv. This allows you to leverage data you’ve already collected while still benefiting from Parcha’s advanced analysis and risk assessment capabilities.
When configured with use_vendor_data_only = true
, the agent will not make fresh API calls to vendor services and will only process the vendor data included in your request.
How It Works
When you provide vendor data:
- The agent skips vendor API calls (no fresh data fetching)
- Uses the vendor data provided in your request
- Still processes all weblinks in the vendor data:
- Scrapes and analyzes article content
- Extracts metadata from articles
- Determines risk factors if not provided
- Augments profiles with extracted information
- Cross-references vendor hits with web presence data
- Generates risk ratings based on the analysis
The use_vendor_data_only
flag only prevents new vendor API calls. All other processing (weblink scraping, LLM analysis, profile augmentation) still occurs.
Vendor Data Specifications
Adverse Media Screening Data
Include adverse media profiles found by your vendor in the adverse_media_event_check
section:
Adverse Media Field Details
Adverse Media Field Details
Required Fields:
id
: Unique identifier for the profilefull_name
: Full name used for matching
Recommended Fields:
associated_countries
: Countries mentioned in adverse mediaassociated_addresses
: Addresses mentioneddate_of_birth
: For age verification (YYYY-MM-DD format)weblinks
: Source articles that will be processed
Optional Fields:
- Name components:
forename
,surname
,middle_name
aliases
: Alternative names- Event details:
title
,summary
,when
,topics
- Demographics:
age_as_of_today
,gender
,occupation
identification_info
: ID documents mentionedbiography_info
: Background information
PEP Screening Data
Include PEP (Politically Exposed Person) matches in the pep_screening_check_v2
section:
PEP Field Details
PEP Field Details
Required Fields:
id
: Unique identifierfull_name
: Full name for matching
Recommended Fields:
pep_type
: Values:domestic
,foreign
,international_org
pep_status
: Values:current
,former
pep_title
: Official titleassociated_countries
: Countries of operationdate_of_birth
: For verificationroles
: Array of political roles with dates
Optional Fields:
- Name components and aliases
- Location information
has_adverse_media
: Boolean flag- Biography and identification details
known_associates
: Associated persons/entitiesrelationships
: Related PEP profiles
Sanctions Screening Data
Include sanctions matches in the sanctions_screening_check_v2
section:
Sanctions Field Details
Sanctions Field Details
Required Fields:
id
: Unique identifierfull_name
: Full name for matching
Recommended Fields:
associated_countries
: Sanctioning countriesdate_of_birth
: For verificationsanction_details
: Array of sanction information with jurisdiction and dates
Optional Fields:
- Name components and aliases
native_character_names
: Names in native script- Location and biographical information
known_associates
: Associated entitiesreports_origin
: Source information
Complete Example Request
Here’s a full example showing how to submit vendor data for all three screening types:
Important Notes
General Guidelines
- Empty Results: Use empty arrays (
[]
) when no matches are found - Date Formats: Use ISO 8601 format (YYYY-MM-DD)
- Country Codes: Use ISO 3166-1 alpha-3 codes (USA, GBR, etc.)
- Topics: Use standardized topic categories (e.g., “financial_crime”, “fraud”, “money_laundering”)
Field Requirements
-
Minimal Required Fields:
id
: Unique identifier (auto-generated if not provided)full_name
: Critical for name matching- Other fields default to empty arrays/null if not provided
-
Critical Fields for Risk Assessment:
is_perpetrator
: For adverse media - indicates if person was perpetrator vs victim/witnessassociated_countries
: For geographic matchingtopics
: For adverse media categorizationdate_of_birth
: For age/identity verificationweblinks
: Source articles that will be scraped and processed
Processing Behavior
When you provide vendor data:
- The agent will still process and analyze all weblinks
- Missing fields like
is_perpetrator
will be determined through analysis - The agent will augment and enrich the vendor data with additional findings
- Cross-referencing with web presence data still occurs for validation
The vendor data integration feature allows you to leverage existing screening data while still benefiting from Parcha’s advanced analysis, cross-referencing, and risk assessment capabilities.