AI/ML System Development for Impersonation Scam Detection

The contract seeks the development and deployment of an AI and machine learning system designed to identify fraudulent communications that impersonate the United States Patent and Trademark Office across multiple channels including email, text messages, phone calls, and physical mail. The system must not only detect deceptive messages mimicking official USPTO outreach but also accurately validate and distinguish legitimate communications from the agency. This requires advanced pattern recognition, natural language analysis, and behavioral modeling to handle the evolving nature of impersonation scams while minimizing false positives and ensuring high fidelity in classification. The work is structured as a subcontract under NAICS code 541512, focusing on custom computer programming services, with a performance location in Alexandria, Virginia, zip code 22314. The opportunity was posted on April 29, 2026, and responses are due by May 6, 2026, at 4:00 PM Eastern Time. The contracting entity is the Department of Commerce, specifically the USPTO, and the project aims to enhance the agency’s ability to protect the public from deceptive schemes while maintaining trust in its official communications. No set-aside provisions or specific small business preferences are indicated, and the work is expected to involve robust data handling, model training, and real-world validation across diverse communication formats.