Extend Launches Verify API to Detect Document Fraud for AI Agents

Extend announced Extend Verify, a document fraud detection API designed for developers and AI agents, according to Help Net Security. The service enables automated systems to assess the integrity of documents before using them to make critical business decisions.

The platform examines visual content, file structure, and revision history to identify suspicious signals. It combines specialized models and forensic checks to flag inputs that may be fraudulent, allowing agents to flag suspicious inputs before acting on them.

How the Technology Works

Extend Verify targets multiple types of document manipulation, including AI-generated forgeries and localized tampering. The system detects changes to specific fields or regions within genuine documents, down to single-character edits. It also identifies conventional digital editing, such as patched text and image overlays.

The tool checks for post-signature changes, providing evidence of alterations made after a document was digitally signed. It returns risk assessments, scores, and traceable evidence to help businesses determine which documents require further human review.

Impact on Business Operations

Businesses rely on documents like bank statements, invoices, and receipts for lending, payments, and insurance claims. Extend CEO Kushal Byatnal stated that moving beyond search-focused agents requires trust in the evidence they act on. The company aims to prevent fraud-related losses and mitigate the difficulty of human review.

AI has made imperceptible forgeries easy to produce, posing risks when fraud escapes review. Extend Verify aims to help developers build document integrity into their pipelines, enabling them to scale up agent deployments while maintaining security.

Performance Metrics

In a challenging benchmark evaluation, Verify accurately detected 96.3 percent of fraudulent documents. The evaluation measured both the identification of fraud and the frequency of false positives on legitimate documents. The tool achieved a 4.0 percent false-positive rate, balancing detection with operational efficiency.

What should you do?

Organizations deploying AI agents for document processing should evaluate tools that verify input integrity. Security teams should monitor for signs of localized tampering and post-signature changes in critical business records. Developers should integrate forensic checks into their pipelines to flag suspicious inputs before automated actions occur.

What to do and how to stay safe: Extend

• Review current AI agent workflows to identify where unverified documents trigger financial or operational actions.
• Implement technical controls that check file structures and revision histories for unexpected modifications.
• Establish procedures for human review when automated systems flag documents with high-risk scores.
• Monitor for emerging threats involving AI-generated forgeries that mimic legitimate business records.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What types of document fraud does Extend Verify detect?

It detects AI-generated documents, localized AI-tampering, conventional digital editing, and post-signature changes.

How accurate is the Extend Verify API?

It detected 96.3 percent of fraudulent documents with a 4.0 percent false-positive rate in benchmark testing.

Who is the target audience for this tool?

Developers and AI agents that need to assess document integrity before making business decisions.