Israel is a significant development center for technology that combats fraud and financial crime. Below are seven of the most promising anti-fraud Israeli startups of 2024. Expect to hear more about them in the future.
Chargeflow
Chargeflow was founded in 2021 by Israeli-American entrepreneurs Ariel Chen and his brother Avia Chen, who had previously assembled the team that founded Babe Cosmetics. They had a successful enterprise but a massive chargeback issue that resulted in the birth of Chargeflow. Chargeflow is the first merchant-centric 100% automated chargeback automation platform. Chargeflow automatically inspects disputes using smart AI algorithms, auto-responding, and keeping its high win rates — it wins > 80% on average!
Chargeflow simplifies the dispute-handling process by allowing AI-driven analysis of disputes and offering a system to collect evidence automatically. This robust platform provides deep insights into dispute trends and risk management, empowering businesses to act proactively and minimize future disputes.
Full integration to all top eCommerce platforms and payment processors results in a frictionless business process, letting businesses focus on growth instead of disputes. Businesses are not only getting their cases resolved sooner, but they are winning them more often as well, reassuring their revenue sources and leading to higher customer confidence.
With an explosive rise in eCommerce transactions, Chargeflow has emerged as a must-have service to protect businesses against wrong chargebacks and fraudulent claims. With a scalable SaaS platform companies can manage their resources effectively, concentrating on their core offerings and enhancing customer satisfaction.
Fugu
Founded by Amir Sadras and Nimrod Shori, Fugu is named after the poisonous and potentially fatal Japanese seafood dish. It tracks payments post-checkout helping online sellers safely accept transactions they would otherwise lose to declines & payment churn amid fears of fraud.
Fugu is the first fraud solution handling risk along the order lifetime including order creation at checkout, post-checkout, pre-shipment, and post-delivery. The company’s multi-tier approach, which leverages machine learning, allows it to validate transactions until shipment, ensuring less transactions are declined due to lack of information.
“Normally, you can optimize [eCommerce transactions] for either conversion or security but not both,” Sadras tells Merchant Fraud Journal. “[However,] by delaying the decision you can decouple payment acceptance from payment verification and optimize for both.”
By decoupling the sales conversion and fraud protection processes, businesses can optimize each component individually, generating substantial benefits by ensuring that both parts function at their best. In practice, this means orders can progress initially, regardless of high-risk indicators. Post-payment verification then follows, providing a detailed examination of the transaction’s risk profile. It provides online businesses with continuous and up-to-date information on the risks inherent to each transaction, allowing decision points such as cancellation before shipment to occur.
“With the slowdown in e-commerce growth and the cost of declining a customer on the rise, e-commerce merchants need to make sure declining a customer is the last resort after all options to accept it were exhausted. “At Fugu, we take potentially fraudulent e-commerce transactions and turn them into delightful revenue. Just like an expert chef takes the poisonous pufferfish and makes a delicious dish,” says Sadras.
Validit.ai
Validit.ai was co-founded by Avivit Yorkevich and Yossi Penias, two professionals with decades of service in the Israeli security services, Validit.ai seeks to turn your smartphone into a professional-grade lie detector machine. The company is seeking to leverage decades of heavy use of polygraphs by the security services for more commercial applications.
“I think that Israel is a polygraph superpower because we use it a lot,” Yossi Penias tells Merchant Fraud Journal.
Lying detectors are based on the observation that for most people lying triggers their fight or flight response, as the prospect of getting caught engages the sympathetic nervous system and leads to a differentiated response in a variety of bioindicators.
The polygraph or the measurement behind the polygraph allows you to identify in a very certain protocol when someone is stressed to a certain degree that when someone is engaging in deception and the polygraph is marked “deception indicated.”
“Because we are dealing with human beings, there isn’t 100% success rate,” says Penias, under the best circumstances and conditions and examiners and equipment and examinees, it’s 85% to 87% accuracy.”
According to Penias, Validit.ai’s product is already basically within the 85% accuracy range that one would expect of a professional-grade lie detector. So how can such a product be used? One major use case Penias was willing to discuss is the processing of insurance claims, where reams of paperwork must be generated and checked by a claims adjuster before the insurer decides whether to make a payout.
In such a case, the value is not in catching the liars, but it’s in giving a green light to people who don’t lie to go through an expedited process that will lead to a faster approval of claims. For people who are flagged by lie detectors, they will go through the normal, lengthy human-driven process that occurs today.
Validit.ai has raised so far $3 million in seed money, including an Israel Innovation Authority grant. They expect to raise a Series A round sometime in mid-2024.
Refine Intelligence
Refine intelligence made the news this past November when it raised $13 million in seed capital for its next generation anti-money laundering solution. Founded by CEO and serial entrepreneur Uri Rivner, alongside COO Oren Kedem and CTO Alon Shacham, the company currently has roughly 20 employees. Refine Intelligence is pioneering the innovative approach of trying to “catch the good guys” instead of the “catching the bad guys” in anti-money laundering. This means profiling legitimate behavior to understand when the activity in bank accounts represents something truly unusual that warrants a suspicious activity report (SAR).
The observation driving the new approach is that with the legacy AML alert providers, between 95% and 98% of alerts are either false positives or nonproductive alerts. “What Refine Intelligence does is that we take the 95% of the population [that is good] and we model that instead,” Rio Miner, Head of Intelligence at the company, told Merchant Fraud Journal. “We baseline what a legitimate person does that generates these 95% false alerts. It’s usually something like a real estate transaction which you do 3-5 times in your entire life and is never something you can model. Or it is you getting a large gift for a special occasion like a bar mitzvah or a wedding.”
By modeling these rare but legitimate events, Refine Intelligence can strip away the noise and then locate the 5% of activity that has an increased likelihood of being criminal.
Although not strictly an anti-fraud startup, Refine Intelligence is bringing a newer approach from the world of eCommerce fraud prevention to anti-money laundering.
“I see a lot of similarities between our approach in the AML landscape and the fraud prevention landscape,” says Daniel Shkedi, Head of Product Marketing at the company. Shkedin previously worked for Israeli anti-fraud companies Forter, Biocatch, and Identiq before joining Refine Intelligence. “Large companies like Forter and Riskified started with looking for the fraud, but now it is all about the enablement, pushing the [credit card] approval rates up and looking at a really small percentage of the overall traffic.” Similarly, Refine Intelligence wants to allow many more legitimate bank transactions to pass through unencumbered and focus on the really small percentage of transactions that are truly problematic.
Redstrings
Redstrings
The solution allows investigators to define what they want to research, and how they are going to investigate it manually and then automate and integrate it into a workflow before moving on to the next case.
The goal is to make Redstrings into a turnkey solution where investigators choose or design their workflow with the platform, and then they pay Redstrings to use the workflow.
“Our vision is to be a home to investigators, the way Quickbooks is for accountants and PhotoShop and Figma are for designers,” says Offer Gombo.
Dtect Vision
Dtect Vision was co-founded by CEO Lior Moyal and two professors, Prof. David Mendlovic, a professor in optics and expert in machine vision, and Prof. Dan Raviv, a world expert in AI and deep neural networks.
Dtect Vision, as its name might imply, is working on technology to detect deepfakes within seconds. The company already has six patents to its name and its main target market is traditional broadcast media and social media.
At its core, Dtect Vision relies on facial recognition technology that differs from what is already available on the market. Most facial recognition technology looks at one picture of your face or a stream of pictures from a video screen that is analyzed frame by frame, Dtect Vision’s technology doesn’t look at the frames themselves, but at the difference of the pixels between one frame to another. In other words, the company’s solution doesn’t look at the information itself but rather a derivative of the information. In addition to its expertise in video analysis, Dtect Vision has added researchers handling still image, audio and text analysis to address all types of fakes made possible by generative AI.
“Tech-wise we believe we are superior to everything out there,” says Moyal. “Now, we’re ready to demo our solution.”
This year is anticipated to be a big year for the budding company, with 50 national elections taking place worldwide, including one in the U.S., meaning a significant increase in disinformation attempts is expected in 2024.
“In the age of generative AI, where anybody can be or can look like any person they want to, you have to put extra emphasis on measures to defend yourself from the negative aspects of that revolution,” says Moyal. “Dtect Vision is a dynamic biometric recognition technology that is key to protecting the world from [unintentional] misinformation and [intentional] disinformation.”
Justt
Justt specializes in automating chargeback mitigation and preventing friendly fraud. Utilizing advanced AI technology, Justt streamlines the chargeback process by integrating seamlessly with over 40 payment service providers (PSPs), enabling real-time evidence collection and response generation. Their system pulls data from over 500 points, such as shipping status and order details, to craft tailored responses for each dispute, significantly improving win rates.
The platform’s fully automated approach eliminates the need for manual intervention, allowing businesses to focus on their core operations. Justt’s AI continuously optimizes its performance through A/B testing and adaptation to industry-specific regulations and challenges. This ensures that companies receive the most effective and efficient fraud prevention and chargeback management available.
Moreover, Justt prioritizes data security, offering enterprise-grade protection for all client information. The company’s solutions cater to various industries, including eCommerce, crypto, travel, and gaming, providing customized strategies to handle unique chargeback and fraud scenarios. This makes Justt a comprehensive and versatile partner in managing financial disputes and protecting revenue streams.