Following the Money: Building Tag to Track Insurance Fraud

Summary

After I left the Marines, I went to work at Pringle, Quinn, Anzano in Belmar, New Jersey, from 2004 to 2010 as a software developer. The firm was best known for representing the Special Investigation Units of insurance carriers, including Allstate New Jersey Insurance Company, Encompass Insurance, Mercury Insurance Group and Selective Insurance Company of America, in challenging complex healthcare insurance fraud schemes. Tag was one of the systems I built to help them do it.

The Firm and the Fight Against Insurance Fraud

Pringle, Quinn, Anzano had offices in Belmar, Morristown, Mount Laurel and Trenton. Its practice covered commercial litigation, governmental relations and regulatory law, municipal law, commercial transactions, real estate and land use, and wills and estates. The part of the practice I worked in was insurance fraud. Those cases involved automobile accident fraud rings, healthcare provider fraud, No Fault PIP, UM and UIM claims, and automobile theft claims.

Cases like that live and die on information. Who is connected to whom, which claims keep showing up, which names come back again and again. That information had to be captured, organized and easy to find for the people building the cases.

Tag: Taming Boxes of Paper Checks

Tag is a system I built at the firm. When the firm went after people for insurance fraud, it would subpoena all of their bank accounts. What came back was boxes and boxes of paper checks, some dating back to the 1980s. The fraud being investigated was mostly by car insurance adjusters, and it involved car insurance companies and medical insurance fraud.

This was before OCR was widespread. The firm had some OCR systems, and they were good with printed text but not great with handwriting. Tag is what I built to handle that flood of paper. It took all of the subpoenaed checks the firm received and extracted all of the information from them into a database, stored by each part of the check.

How the Workflow Ran

1. Batches: data entry people received batches of scanned checks. Each batch was locked to one person at a time.
2. Front, back and fields: the front and back of each check showed on screen, with the fields on the right.
3. OCR first: OCR pre-populated most of the fields. It was highly accurate on printed names and addresses.
4. People for handwriting: every part of the check was entered into its field, and handwritten fields still needed manual review.
5. QA: a separate quality assurance workflow followed data entry.

OCR did the printed fields and people handled the handwriting. Together that sped up throughput and improved accuracy.

Following the Money

Attorneys and paralegals could search the database, for example to match signatures across checks. It was a track the money approach to building fraud evidence, and it found connections between people the attorneys had no idea about.

One case stands out. Two men supposedly were not connected, and the attorneys could find no real ties between them. Then the checks showed that one had given the other's son a wedding gift of several hundred thousand dollars. That pointed to a payoff.

Tag worked hand in hand with SIU, which tracked the people on the firm's radar and the connections between them. SIU showed who kept turning up together. Tag showed where the money went.

What the Job Covered

I owned three things: web application development, database administration and data analysis. I wrote the applications, ran the databases behind them and dug into the data to find what mattered. That mix carried straight over from my time in the Marines, where I built web applications and relational databases for the Marine Corps Logistics Command.

What It Taught Me

Building software for a legal team taught me that the users are the experts. Lawyers and investigators know their cases. My job was to give them tools that fit how they already worked and to keep the data accurate, because in litigation a wrong record is worse than no record. I still build that way: start with the people who will use the system every day and make the data trustworthy.

The Other Systems I Built There

• SIU, the app that tracked people on the firm's radar for insurance fraud
• AutoArb, the automated workflow system for PIP arbitration litigation
• The technologies and experience behind all three systems

In My Words

When a stack of old paper checks showed that one man had given another man's son a wedding gift worth several hundred thousand dollars, I knew following the money in a database could uncover what nobody would admit.

Brian Messenlehner

Project leadership by Brian Messenlehner.

Need a Custom App for Your Legal Team?

From case tracking databases to secure web and mobile apps for law firms and investigators, I build software that fits how your team works. Tell me what you need.