In this week’s edition of Finovate Global, we feature Uri Rivner, co-founder and CEO of Refine Intelligence. The Tel Aviv, Israel-based company, founded in 2022, made its Finovate debut earlier this year at FinovateEurope. At the conference, Refine Intelligence demoed its technology, Life Story Analytics, that leverages AI to help banks better defend themselves against money laundering.

We discussed the challenge of fighting financial crime, the innovations that Refine Intelligence brings to the market, and the relationship between upstarts and incumbents in Israel’s dynamic, fintech and financial services ecosystem.

What problem does Refine Intelligence solve and who does it solve it for?

Uri Rivner: If you’re a bank, your AML Operations team is massive, and needs to grow every year to cope with growing alert volume. But the team can have a pretty frustrating daily routine, as almost all the alerts they’re investigating end up being totally legit activities done by the customer. 

Take an account that did a large wire transfer to Mexico for the first time. The AML Transaction Monitoring is screaming like a banshee – maybe there’s money laundering here? But after investigating, the team finds out the customer just has a daughter studying in Mexico, and this was to pay her tuition. 

Years ago banks knew these life stories, because everything was done at the branch. But now with digital transformation, banks have lost that superpower.

At Refine intelligence, our mission is to help banks regain that superpower of really knowing their customers’ life stories, so their financial crime teams can quickly clear AML or scam alerts triggered by legitimate customer activity. We work with Risk, Financial Crime, BSA and AML teams. Fraud teams look at our technology to help with scam operations.

How does Refine Intelligence solve this problem better than other companies?

Rivner: Refine Intelligence takes a unique approach for fighting Financial Crime – we call it ‘Catching the Good Guys.’ 

Think of someone who got married and now deposits a large amount of cash from wedding gifts. Or a couple withdrawing cash in order to pay for a big renovation project. Think of people starting a new cash-intensive job, or depositing money from a fundraiser. These are all legitimate activities that look abnormal, triggering transaction monitoring alerts. 

Refine discovers these sort of “life stories,” i.e. legit customer activities behind a flagged anomaly. There are two ways to do that:

The first is to ask the customer and Refine provides that capability through our Digital User Outreach which allows a bank to reach out to customers automatically and collect their explanation within minutes.

The second way is to train AI to recognize the life story behind an anomaly, without reaching out to the customer. Our Life Story Analytics does that, and the training uses our unique, proprietary data set of genuine explanations.

The outcome: clear, fast evidence that helps AML teams clear away falsely flagged anomalies by identifying the legitimate customer activity behind them. 

Who are Refine Intelligence’s primary customers? How do you reach them?

Rivner: We work with large to mid-sized banks who operate a big team of investigators to look into AML alerts. Refine helps those banks reduce their operational effort significantly without making any change in the Transaction Monitoring system. 

Our founders and senior management team have been working with financial crime units for decades, and we expand our reach via participating in events such as Finovate, as well as our own virtual events. 

Can you tell us about a favorite implementation or deployment of your technology?

Rivner: A Top 50 bank in the U.S. deployed Refine Intelligence to handle customer outreach for AML. Before using Refine, the AML team approached the branch when they couldn’t find a good explanation to a flagged anomaly. The branch tried calling customers, leaving messages and chasing them for answers. A district manager described the situation as “we are the punching bag of the AML team.”

After the bank started working with Refine, it became clear why the existing RFI (Request for Information) process was driving everyone crazy. The average time to complete a customer outreach was 16 days with 3.6 back-and-forth emails between the AML team and the branch, as initial responses were often insufficient. The process consumed resources that were better used elsewhere.

Refine Digital User Outreach automated the process by messaging customers through digital channels. Response time was cut from two weeks to two minutes, completely changing the game for the Operations team who could work on alerts without interruption, receiving quality responses. With an 85% answer rate, the digital process outperformed manual outreach. Data collected was structured and allowed analysis and benchmarking, and soon most RFIs (Requests for Information) turned digital using the Refine system. The AML team loves the new approach.

What in your background gave you the confidence to respond to this challenge?

Rivner: I’ve been fighting online fraud for 20 years in Cyota, RSA and BioCatch – which I co-founded. This helped me take an outside look at the way AML was operating and realize that the current paradigm isn’t sustainable.

Online fraud detection benefits from context-rich signals that go well beyond transaction monitoring, device analysis, geo-location or behavioral biometrics. These signals feed into AI that is trained using a huge pool of fraud cases, as victims report fraud in their bank account. But no one reports money laundering in their own account, and when a bank files a Suspicious Activity Report, they never get feedback from authorities. You can’t train AI to recognize bad guys without feedback, so the industry had to revert to anomaly detection. 

You can get more and more efficient in anomaly detection, but at the end of the day most of what you find is irregular activity in good people’s accounts. Any improvement in detecting bad guys is doomed to be marginal. And that’s not good – the industry needs a game changer… 

This brought the insight of reversing the focus, to “Catching the good guys,” that is, detecting the legit activities that were falsely flagged as anomalies.

What is the fintech industry like in Israel? What is the relationship between fintech startups and the country’s established financial services sector?

Rivner: Israel, widely known as the ‘start up nation,’ is a powerhouse of cyber, fintech and financial crime fighting. Many market-shaping startups grew up in Israel: Cyota, now RSA Outseer, was first to introduce risk-based authentication using device and geo-location analysis. IBM Trusteer was first to launch an anti-Trojan tech. BioCatch was first to leverage behavioral biometrics for online fraud and scam detection. Forter and Riskified pioneered the chargeback guarantee market in eCommerce, Simplex did the same in crypto, and DoubleVerify prevents fraud in the digital advertising market. The largest global player in AML is Nice Actimize, and companies like EverC and ThetaRay help acquiring banks and payment providers manage financial crime risks. This might explain why there’s a vibrant community of fraud fighters in Tel Aviv.

Interestingly enough, the local Israeli market has never been a big target for those innovators. Most Israeli Fintech startups work directly with global design partners, who recognize the disruptive, out-of-the-box thinking behind their technology.

You recently demoed your technology at FinovateEurope. What was that experience like?

Rivner: Demoing at FinovateEurope was fantastic! We were thrilled to have the opportunity to demo together with so many other innovative fintech companies, and to meet with banks that are looking to incorporate innovative technologies into their operations. The experience was very TED-like, professional, and the vibe was exciting.

What are your goals for Refine Intelligence? What can we expect from the company over the balance of 2023?

Rivner: Everyone we talk to is very excited about what Refine is doing. When showing our Digital Outreach capabilities, AML teams come up with so many ways to use it effectively – from automating requests of information for resolving transaction monitoring alerts to helping the line of business with Enhanced Due Diligence and Cash Structuring education. Fraud teams are particularly interested in digital outreach to potential scam victims, and it is a great way to conduct rapid investigations of incoming wire and check deposits.

But the biggest amount of interest is in our other bit of magic – Life Story Analytics. That’s where we train AI to recognize the legit customer activity behind a flagged anomaly, without reaching out to the user. Financial Crime teams are excited about the notion of keeping their AML transaction monitoring or scam detection models as is, despite the high degree of false positives, and letting AI sweep aside the legit customer activities so what’s left are the real unexplained anomalies that might be money laundering or scam victims. That’s going to be a major area of expansion for Refine.

Photo by Haley Black