To Combat Billions In Unemployment Benefit Fraud, Startup SentiLink Raises $70 Million

Harris’ team has seen a huge uptick in fraud rates affecting their clients, as high as 90% among new applications, associated with the CARES Act COVID relief. Fraudsters have been using the same name, social security number or date of birth in several applications, filing in high volumes in several states.

According to Harris, his team is currently verifying around a million account openings per day, and is working with more than 100 financial institutions – due to a non-disclosure agreement Harris could not comment on which financial institutions his company serves.

The company says that beyond simply using artificial intelligence to detect fraud, they have a risk operations team that catches in real time cases of synthetic fraud – a form of identity theft in which the defrauder combines a stolen Social Security Number (SSN) and fake information to create a false identity – that would normally go unnoticed by their clients.

Harris discovered this type of fraud when he was working as a data scientist at Affirm in 2017. At the time, synthetic fraud was relatively unknown, so when he saw that crooks were creating brand new identities instead of stealing  existing ones to apply for credit, he founded SentiLink to focus on tackling this new scam. “We realized this was a really big issue and that nobody in the financial services industry was talking about it,” says Harris.

Now, criminals are creating new identities or stealing existing ones to tap into unemployment benefits. Harris says the problem is not only them stealing from the government, but uncovering the tactics they use to deposit the stolen funds.

“What a lot of people don’t realize is that as a fraudster you have to be able to use the money stolen, and put it into the financial system,” Harris says.

To Harris, the biggest differentiation in SentiLink’s approach is how much it emphasizes “deep understanding of fraud and identity in our models.”

“We have a team of fraud investigators that manually review applications every day looking for fraud, and we use their insights and discoveries in our fraud models and technology,” he told TechCrunch. “This deep understanding is so important to us that every Friday the entire company spends an hour reviewing fraud cases.”

SentiLink, Harris added, focuses on “deeply” understanding fraud and identity, and then using technology to productionalize these insights.  Those discoveries include the deterioration of phone/name match data and uncovering “same name” fraud. “This deep understanding is so important that SentiLink employs a team of risk analysts whose full time job is to investigate new kinds of fraud and discover what the fraudsters are doing,” the company says.

SentiLink, like so many other startups, saw an increase in business during the COVID-19 pandemic. “The various government assistance programs were rife with fraud. This had a cascading effect throughout financial services, where fraudsters that had successfully stolen government money attempted to launder it into the financial system,” Harris said. “As a result we’ve been very busy, particularly with checking and savings accounts that until now have had relatively little fraud.”

The startup plans to use its new capital to build out its product suite and do some hiring. Today it has 25 employees, with five accepted offers, and expects to end the year with a headcount of 45-50.