nSure.ai is a world leader in AI-powered advanced fraud prevention for high-risk digital transactions, as the company provides a unique multi-tenant platform that empowers customers with their own AI/Machine learning model. Pulse 2.0 interviewed nSure.ai CEO and co-founder Alex Zeltcer to learn more.
Alex Zeltcer’s Background
Alex Zeltcer spent half of his career in large enterprises and the other half in startups. “I started my career at Comverse, a large software provider for the telecom industry. I ran one of their business lines and had about 250-300 people working in the division. From there, I went to Dassault Systemes, a large French software firm, where I ran a company they had acquired. My team grew our sales from $30 million to $100 million under my direction. Next, I began the startup phase of my career. I worked with a team of people from Dassault, and together we created a new company called Artizone. It was an online farmers market that delivered groceries from small farms to customers’ homes in Dallas, Chicago, and Denver. Artizone was eventually sold to a large grocery chain in the Midwest, and I returned to Israel to work for a gift card marketplace startup called Zeek,” said Zeltcer. “It was at Zeek where we first envisioned the idea behind nSure.ai. When I started there, we were processing about $3 million to $4 million in annual sales and scaled it to about $100 million a year within 18 months. While there, we realized that the technology we developed to fight against fraud was probably more valuable than the entire marketplace, so we decided to split off the fraud prevention unit. We brought in investors to acquire the technology and the team and founded nSure.ai in March 2019.”
Formation Of nSure.ai
How did the idea for nSure.ai come together? Zeltcer replied:
My first week at Zeek started great. We saw a huge 40% spike in sales, which meant we had sold all our inventory. We were really happy until we realized that 40% of our sales were fraudulent – and we were victims of players who had used stolen credit and credentials. We were then hit with chargebacks for that 40% of sales. Given our margin on this, which was almost nothing, it was very painful to be hit with that level of fraud.
We shut down operations and decided to use a number of rule-based systems combined with a number of databases. Ziv, my co-founder of nSure.ai, decided around the beginning of 2014 to invest heavily in building machine learning models that could detect fraud more effectively because there was no solution in the market that could support our needs.
It took us some time to grow the business, and it was generating a few million dollars a year. At that point, gross margin wasn’t very important to us. However, as the company grew to 50 million in sales and then a year later it was at $100-$120 million in sales we realized that fraud was still eating away a large chunk of our margin.
We were declining a lot of transactions to prevent fraud, which had the effect of driving our customer acquisition costs very high. Our fraud detection model was declining about 20% of all transactions, and the payment processors were declining another 10-15%, which meant we declined 35% of all customer transactions.
When we analyzed the numbers further, the decline rate on new customers was about 50%, because new shoppers are riskier than shoppers with a track record. If you decline 50% of new shoppers, it means your customer acquisition costs are doubled, and we were not showing a profit on paper. It was clear we couldn’t afford that level of decline rate.
When we built our machine learning system, we were actually able to reduce the decline rate from 20% to just 2%. We were able to get the payment processor down as well, and soon we were declining about 5% of new shoppers, which is a much different ballgame.
We realized we were only able to achieve 20% using tools like Forter and Riskified, which were the best tools on the market at the time. Our tool was fundamentally different and could be helpful for other companies dealing with digital products that were prime targets for fraudsters. And at that point, we decided to spin off the fraud detection portion of the business.
Challenges Faced
What were some of the challenges that Zeltcer faced in building nSure.ai? Zeltcer acknowledged:
Like any startup, the first thing you have to do is prove your solution works. We had a head start because of the technology and team we acquired. However, proving it worked was not a walk in the park. The first time we launched, a gift card vendor was hit with a lot of fraud. It took us a few months to identify what was wrong with the model and fix it, and we were able to stabilize the product.
The next step is finding enough clients, to prove what we call in the startup world product-market fit. That led to finding an issue in our sales cycle, because getting clients to believe our story is pretty tough. When you look at it from their perspective, they are declining 20-25% of their shoppers, and we’re telling them we can approve 98% of transactions, and that there are only 2% of transactions that are fraud. Essentially, they are hearing us tell them they aren’t doing a good job at their job.
We also charge more than our competitors so on the surface, sans the business case, it’s a tougher hill to climb. Many of our potential customers within organizations heard our pitch and felt that if we were right, they might look bad.
We had to adjust our tactics, and used client success stories to gain credibility. Our ROI charts showed that even with our higher price, our customers could easily justify the investment because the return was significantly better. Lastly, we helped our customers understand that we weren’t taking away their job, we were really making them heroes. Our value proposition is such that it helps position them within the organization as the team that was not only fighting fraud but was directly responsible for increasing revenue. That took us about 18 months to get right.
We had some challenges with fundraising, which is pretty typical of a startup. Today, our biggest challenge is scaling. We’ve had some growing pains, but at the same time, we’d like to double our headcount over the next year and a half. That type of growth requires developing an infrastructure within the organization to handle things like communication, organizational alignment, and defining responsibilities, and that’s what we are working hard on now.
Core Products:
What are nSure.ai’s core products and features? The company’s core product is liability shifting for merchants that sell digital goods, services, and assets. And nSure.ai works primarily with four industries today: prepaid gift cards, gaming, cryptocurrency, and travel.
“We provide them with a decision for every transaction, and charge them per transaction. If an approved transaction turns out to be fraudulent, we pay back the charge,” Zeltcer explained. “In addition to the core service, we have several add-on services. We have a service where we cover fraud-related chargebacks and service-related chargebacks. We have several services that take offensive action against fraudsters, inflicting a price on criminals who try using stolen credit cards. Lastly, we offer a service that allows merchants to enter a fraud monitoring program by Visa or Mastercard. If the fraud goes beyond a certain threshold, we take responsibility for the penalties.”
Evolution Of nSure.ai’s Technology
How has nSure.ai’s technology evolved since you’ve launched? “At first, we had one AI model. We knew up front that we needed models that were customized per client to achieve the level of accuracy that we offer. So, we had to build an infrastructure that automates the entire data science pipeline to be scalable, which was not something simple and required significant investment on our part,” noted Zeltcer. “We’ve also gone from working with gift card companies to service the gaming, crypto, and travel industries, and each of them needed different form patterns and elements to support those verticals. We’re starting to reach the edge of our platform’s capabilities from a data science perspective, as we’re processing about 4 billion transactions a year. So there are a lot of infrastructure enhancements we are building that need to be implemented in a scalable manner without breaking the bank.”
Significant Milestones
What are some of nSure.ai’s most significant milestones? “Obviously establishing the company and getting our first client were our first two milestones. We had a fraudulent client, which is a milestone of sorts,” Zeltcer pointed out. “The next big milestone was adding the gaming industry to the verticals we covered. We decided to target that market during COVID when gaming was growing faster than most other industries. Our next milestone was probably realizing that we had a market fit with that industry, and reaching a million dollars in revenue. We’ve had some fundraising milestones and sales milestones since and have started to win industry awards, as well as being granted another patent.”
Customer Success Stories
Can you share any specific customer success stories? “InComm Payments is the biggest client we have today. They sell billions in gift cards a year, some of which are prepaid debit cards for Visa, Mastercard, and AmEx. We helped them increase their sales by more than 10%, bringing their approval rate from 80% to 90%. We also took complete responsibility for risk, so that was something they didn’t have to think about. InComm used to have a manual review team, and we eliminated the need for that review, which was quite significant. We are working with a gaming platform that I can’t name here. We decreased their decline rate by more than 50% and helped approve roughly 1,000 new shoppers that would have been rejected every day. That’s a huge value for any business,” Zeltcer revealed. “We are also working with a crypto exchange and helped grow their revenue by 50% by reducing their need to conduct KYC. KYC isn’t required for most transactions, because they are below a threshold. However, most crypto exchanges require it as a way to fight against fraud. This creates a lot of onboarding friction, because customers have to share a picture of themselves with a driver’s license, and the reality is it’s fairly ineffective in the fight against fraud. When you look at all three of these customers, it’s about enabling them to generate incremental value and safely approve more transactions.”
Revenue Growth
When I asked Zeltcer about revenue, he cited 280% growth from 2021 to 2022 and expects to see similar growth this year.
Total Addressable Market
What is the total addressable market nSure.ai is pursuing? “The combination of digital goods and services, where prepaid gaming and travel is, is about a trillion dollars a year in total processed volume. If you look at crypto, that’s about somewhere between half a trillion and a trillion,” Zeltcer assessed. “And if you look at FinTech, which is the next industry we’re targeting, that’s another half a trillion to a trillion dollars. The total volume is about $3 trillion, and our take rate today is roughly 1%. So, we are going after $20-$30 billion of the total addressable market.”
Differentiation From The Competition
What differentiates nSure.ai from its competitors? “The main thing is our focus on the digital industry and preventing scalable fraud. If you look at our competitors, they came from e-commerce and are trying to adapt lessons from that payment ecosystem to the digital goods market,” Zeltcer affirmed. “We took a completely different approach. Because of our authentic experience as a merchant, we bring to the table something no one else has – the merchant’s perspective – and a dedicated AI model that is designed for these types of businesses.”
Future Company Goals
What are some of nSure.ai’s future company goals? “The main thing we want to do is build a sustainable business and provide a highly reliable, quality service to our customers. It’s a white glove offering designed for large enterprises. As a result, our goal is to be uniquely customer-focused, to be viewed by our customers as a trusted partner they don’t want to live without and to enable them to grow their business as fast and as profitably as they can,” Zeltcer concluded.
