Case Study
$30,000 in fraud losses. Caught and stopped.
How Providence AI built custom fraud detection and operational intelligence systems for a mid-market American manufacturer serving customers from individual consumers to NASA and every branch of the U.S. military.
The Client
American Flags Express is an American-made flag and flagpole manufacturer based in Wisconsin, serving over 300,000 newsletter subscribers and customers ranging from individual patriots to every branch of the U.S. military, NASA, and major corporations. The company operates a proprietary Half-Staff Alert System that notifies subscribers across all 50 states when state governors or the U.S. President declare half-staff.
- Industry
- American-made flags, accessories, and flagpoles
- Customers
- Individual consumers, U.S. military, NASA, corporations, government
- Subscribers
- 300,000+
- Staff
- 7–8 full-time, 30+ seasonal
The Problem
American Flags Express lost approximately $30,000 to fraudulent orders in the first four months of 2026 alone — annualizing to roughly $90,000/year in exposure. A small full-time team was stretched thin running operations, leaving little capacity to manually review every order for fraud signals. High-value fraudulent orders were slipping through to shipment, then resulting in chargebacks and lost inventory.
Beyond fraud, the company faced operational blind spots common to growing businesses: no systematic competitive intelligence, no data-driven purchasing decisions, and limited time for strategic R&D on new product development.
The Approach
Rather than recommending an enterprise fraud platform that would cost more than the losses themselves, Providence AI built a custom system tailored to the company's actual operations and infrastructure. Critically, the fraud detection system was designed to operate outside the client's network — an architectural decision that addressed the president's deep concerns about cyber security, having experienced past intrusions.
Over several weeks, we built four complementary systems:
- 1
Fraud Detection System
Order confirmation emails are automatically forwarded to an isolated email environment, where an AI agent evaluates each order against custom risk criteria, assigns a risk score, and immediately alerts management to high-risk transactions before fulfillment.
- 2
R&D Intelligence Agent
A weekly automated competitive analysis system that monitors the company's top 10 competitors, identifies product line gaps, surfaces market trends, and delivers a CEO-ready strategic brief every Monday morning.
- 3
Inventory & Trends Agent
Connected to the fraud detection system's order log, this agent compiles every order placed throughout the week and delivers a data-driven purchasing report to the warehouse manager every Friday night, enabling trend-based inventory decisions.
- 4
Personal Productivity Infrastructure
A custom workflow combining Telegram, Notion, and AI agents to centralize task management, habit tracking, and goal alignment — built first as proof of methodology before deployment to other systems.
The Outcome
All four systems are operational and running in production. The fraud detection system has processed over 100 orders, flagging high-risk transactions for review. The R&D Intelligence Agent has delivered multiple weekly competitive briefs to company leadership. The Inventory & Trends Agent began running automated weekly reports during Memorial Day weekend — the company's busiest season.
- —$90,000+/year in fraud exposure addressed
- —8–12 hours/week of competitive research replaced with a 15-minute Monday brief
- —Data-driven purchasing decisions replacing intuition-based ordering
- —Operational visibility increased without adding headcount
"[Testimonial pending approval]"
What This Means For Your Business
The American Flags Express engagement validated something important: small and mid-market businesses don't need enterprise AI platforms. They need fitted systems, built around their actual operations, designed to be maintained by their actual teams. The patterns we developed here — fraud detection, intelligence agents, inventory optimization — translate directly to healthcare practices, trades businesses, and specialty retailers facing similar operational challenges.