Skip to main content

​
Product Policy Engine Overview

The Product Policy Engine is an agentic workflow that autonomously analyzes product listings against policy guidelines. It intelligently decides which tools to use (web research, recall databases, etc.) based on the product type and automatically gathers evidence to make informed compliance decisions.

​
Engine Inputs

The Product Policy Engine Configuration has the following parameters:
  • product_listing: required. The product listing text/content to analyze for policy compliance. Can include text descriptions, URLs, or any product information.
  • product_images: optional. Image URLs to analyze for policy compliance (comma/newline/whitespace separated).
  • policy_version_id: required. The Policy version ID containing the compliance rules to evaluate against.
See Template Strings for dynamic parameter configuration.

​
Engine Output

The output includes a comprehensive compliance investigation result: 
{
  "policy_applied": {
    "name": "Product Policy Name",
    "version": "policy-version-id"
  },
  "compliance_result": {
    "verdict": "APPROVED | DENIED | FURTHER_REVIEW_NEEDED",
    "summary": "Executive summary of the compliance analysis with key findings",
    "summary_evidences": [
      {
        "reference": "$REF_1$",
        "evidence_type": "web_link | screenshot | artifact",
        "evidence_data": "URL or artifact reference",
        "evidence_name": "Short name",
        "evidence_description": "Why this evidence is relevant"
      }
    ],
    "investigation_findings": [
      {
        "rule_title": "Policy Rule Title",
        "rule_description": "Rule description from policy",
        "hit": true,
        "reasoning": "Explanation of why this rule was triggered",
        "evidence_references": ["$REF_1$", "$REF_2$"]
      }
    ]
  }
}

​
Example Usage

1

Create a Policy

First, create a Policy in the Policies section with your compliance rules and guidelines.
2

Create an Agent

Click on the “Add Agent” button in the top right corner of the Agents page. Enter a name and an optional description of your Agent.
3

Select the Product Policy Engine

4

Configure the engine

$ starts a template string
  • product_listing: $product_listing
  • product_images: $product_images
  • policy_version_id: (select your policy version)
5

Create the Agent

Hit the Create button.
6

Run a job

Create a new job and provide the product listing content to analyze.

​
How It Works

The Product Policy Engine is an agentic workflow that autonomously:
  1. Analyzes the product listing - Extracts key information from text and images
  2. Decides which tools to use - Based on product type, may check recall databases, perform web research, or gather regulatory information
  3. Evaluates against policy rules - Systematically checks each rule in your policy guidelines
  4. Collects evidence - Gathers web links, screenshots, and data to support findings
  5. Makes compliance determination - Provides a verdict with reasoning and evidence references

​
Use Cases

  • Marketplace compliance: Ensure product listings meet marketplace policies before approval
  • Regulatory compliance: Automatically check products against safety regulations
  • Content moderation: Evaluate listings for prohibited content or misleading claims
  • Quality assurance: Assess product descriptions for accuracy and completeness