We’re excited to announce the launch of our new Fraud Detection API!

This latest addition to our suite of AI-powered security tools can be used to detect fabricated, manipulated, and AI-generated content in documents.

Like all Cloudmersive security APIs, the Fraud Detection API is easy to integrate, easy to use, and able to significantly improve your threat profile.

Why Fraud Detection matters in document workflows

Spotting fraudulent documents in a crowd has always been a difficult security challenge. Sophisticated threat actors have always found ways to carefully mimic legitimate documents and bypass both manual and automated document reviews.

With the advent of commercial GenAI, however, the scale of that security challenge has skyrocketed. Jail-broken GenAI models have become popular in the black market, and they’ve made it much easier for less sophisticated threat actors to generate convincing fake documents and launch larger scale fraud campaigns. This has increased the amount of fraud we can expect to see in our document processing workflows, and it’s necessitated larger-scale fraud detection solutions.

Cloudmersive Fraud Detection is designed to combat fraud at scale on behalf of your organization.

What the Fraud Detection API Does

Let’s take a closer look at how the Fraud Detection API operates.

Breaking down Fraud Detection analysis

The Fraud Detection API analyzes document contents and returns a risk score and a clean result diagnosis. The risk score (0.0 – 1.0) describes the likelihood (i.e., risk) that the document contains some form of fraudulent content. Scores at the higher end of the risk score range are high risk fraud, and they return a clean result: false diagnosis. Scores at the lower end of the risk score range are low risk fraud, and they return a clean result: true diagnosis.

Risk scores are accompanied with an analysis rationale. This rationale is a natural language description justifying how the underlying Fraud Detection model arrived at its risk score and clean result conclusions.

Additional context for Fraud Detection analysis

In addition to the risk score, clean result diagnosis, and analysis rationale, the Fraud Detection API offers several document classification indicators. These indicators (listed below) are largely descriptive and neutral, identifying what types of clauses or elements are present within the scanned document.

• Contains Financial Liability
• Contains Sensitive Information Collection
• Contains Asset Transfer
• Contains Purchase Agreement
• Contains Employment Agreement
• Contains Expired Document
• Contains AI Generated Content

A Contains Financial Liability document, for example, is not inherently risky; however, a high risk score accompanied by a Contains Financial Liability response would indicate a significant financial security threat to your business.

Integration with the broader Cloudmersive platform

The Fraud Detection API is designed to work seamlessly alongside other Cloudmersive APIs in both new and existing automated workflows.

Fraud Detection fits naturally in workflows involving Document AI processing actions. You can, for example, call the Fraud Detection API prior to calling the Field Extraction API to avoid parsing data from an illegitimate document.

Fraud Detection also fits well in any document security workflow at the point of document upload. Fraudulent documents will most often pass Virus Scan API checks, so it’s a good idea to implement a Fraud Detection step after the Virus Scan API finishes scanning.

Finally, Fraud Detection is included in Cloudmersive AI Email Protect, which comprehensively blocks threats entering your organization’s inbox. AI Email Protect looks for viruses, malware, fraud, phishing, spam, and NSFW content simultaneously.

Learn More

If you’re interested in learning more about the Fraud Detection API, please visit our Fraud Detection home page and also visit the Fraud Detection swagger page for ready-to-run code examples.

If you have any additional questions, please reach out to a member of our sales team for expert advice.