Why don’t more fraudsters go to jail?

What does it take to catch a fraudster in the act — forging signatures, handing over fake documents, signing for a car they were never going to pay for — and still not be certain they'll face consequences?

That's the question running through this episode of Good Question, where I sat down with fraud detective Marc Evans alongside our frequent co-host, Ronan Burke, CEO and co-founder of Inscribe. Marc is the founder of Fraud Hero and one of the fraud community's most vocal educators. His origin story says a lot: he nearly lost his shot at becoming a police officer after discovering, during his own background check, that someone had opened a fraudulent account in his name. His first fraud case was, in a sense, his own.

The conversation moves from the street-level mechanics of document fraud to the systemic failures that let criminals walk, and lands somewhere that shapes everything Marc does: by the time law enforcement has a case, it's usually already too late.

Fraud is never just one crime

One of Marc's early cases started as a simple check fraud report. It ended up touching mail theft, auto theft, identity theft, document fraud, forgery, and organized crime, all feeding into each other.

The criminals were stealing mail daily to harvest checks, then creating fake DMV temporary licenses and fraudulent bank statements to use at car dealerships. The cars became the vehicles they used to continue the theft ring. The whole operation was self-sustaining.

For investigators, and for fraud teams at banks and fintechs, that complexity matters. What looks like a single fraudulent document at the point of review may be one piece of a much larger picture. The document is a clue, not the whole story.

First-party fraud is frequently misidentified

Marc pushed back on a common assumption in the industry: that first-party fraud is straightforward to identify. In his experience, the opposite is often true.

He described cases where genuine victims of account takeover or identity theft get flagged by their own financial institution as first-party fraudsters, because a two-factor authentication appeared in the logs, or because the transaction looked like it came from the account holder. Sometimes the real explanation is social engineering. Other times, the customer was an unknowing money mule, their account used to move funds without their knowledge.

Getting that call wrong has serious consequences. A real victim denied a claim loses money they may never recover. A wrongly flagged customer can face account closure, damaged credit, or worse. Marc has had to intervene on behalf of customers who were days away from prosecution for fraud they didn't commit.

The takeaway for investigators: first-party fraud exists, but the bar for making that determination should be high, and the methodology needs to account for how sophisticated account takeover and social engineering have become.

Document fraud: from crude forgeries to AI-assisted counterfeits

Marc spent six years as a financial crimes detective, and in that time he watched document fraud evolve significantly. Early on, fake DMV temporary permits were often easy to spot: wrong governor's name, expired templates, incorrect fonts, license numbers that didn't conform to state formatting.

He described one memorable sting at a car dealership where the suspect was caught mid-process, calmly signing forged documents in front of undercover officers. The tell? The temporary driving permit used a template from two governors ago.

Today, those easy tells are disappearing. Marc demonstrated live during our conversation how consumer AI tools can modify documents with a single text prompt, changing names, amounts, and dates in seconds with no design skill required. What used to take Photoshop expertise and practice now takes a sentence.

Ronan added context from Inscribe's data: one in 16 documents submitted through digital channels has been tampered with, and AI-assisted document alteration has increased 200% year over year. Beyond static documents, fraudsters are now using LLMs to generate real-time phone scripts to defeat security questions, and fabricating spending patterns to make fraudulent accounts look legitimate.

Why prosecution is harder than it looks

The question in the episode title has a frustratingly structural answer.

Once a detective completes an investigation and submits a case, it moves out of their hands entirely. It becomes the DA's responsibility to prosecute, and the court's to decide. At each step, the case can stall or collapse: a DA unfamiliar with fraud mechanics who can't explain the crime to a judge; a threshold dollar amount that classifies the offense as a misdemeanor; a suspect in another jurisdiction or country; a bank that requires a subpoena before sharing even the name of a receiving institution, by which point the funds have moved into cryptocurrency and vanished.

Marc described a business email compromise case where he could have recovered the wired funds with a single phone call from the receiving bank — just a freeze, pending paperwork. The bank declined. The money was gone within weeks. The victim did everything right.

The deterrence problem makes this worse. When fraud carries low penalties relative to potential gain, when an organized ring can net hundreds of thousands of dollars and face probation, the risk-reward calculus favors continued operation. Marc was direct: tougher, more consistent prosecution would change that math. Until then, the asymmetry benefits criminals.

The case for prevention, and what law enforcement actually needs

The insight Marc most wants fraud fighters to internalize is a reframe of the whole problem. Most of the industry's energy, its reporting structures, recovery protocols, and bank-to-law-enforcement workflows, is oriented around responding to fraud that has already happened. Marc built Fraud Hero because, as a detective, he kept arriving too late.

Prevention requires different questions. Not "how do we recover from this?" but "how do we stop this from reaching us in the first place?" For banks and fintechs, that means detection infrastructure that catches manipulation before funds move. For law enforcement, it means better tools that simplify the articulation of fraud, because writing a 61-page arrest report explaining why a document is fake is not a sustainable workflow.

On that point, Marc was specific about what he needs from technology vendors: plain-language explainability. When a fraud detection tool produces a clear, jargon-free summary of why a document is fraudulent, not just a score but a reason, it makes every downstream step easier. Investigators write better reports. DAs can explain the crime to judges. Juries understand what happened. The evidentiary chain holds.