For years, cryptocurrency exchanges were regarded as the Wild West of finance. For fraudsters, they represented something even better. “Fraud heaven” is how Pascal Podvin, co-founder of nSure.ai, describes the environment that helped shape his company’s approach to fighting online crime.
Crypto exchanges, digital gift cards and online gaming platforms share a common characteristic. Transactions happen instantly, funds move quickly and, once completed, are often impossible to recover. For criminals, the rewards can be immediate and the risks relatively low. “Imagine you can steal and monetise what you stole within one second,” says Podvin. “That’s exactly what crypto offers.”
Yet while cryptocurrency is often criticised for attracting fraudsters, Podvin believes it has also become an invaluable testing ground. The fraud techniques that first emerged in crypto are increasingly appearing in mainstream banking and digital commerce. What happens in crypto today will arrive in traditional finance tomorrow.
That perspective has led Podvin to challenge one of the fundamental assumptions underpinning the fraud prevention industry. Most fraud solutions are designed to reduce fraud by declining suspicious transactions. The difficulty is that every declined transaction represents potential lost revenue.
“You can have zero fraud,” he says, “but you will have very little revenue.”
It is a deceptively simple observation, but one that sits at the heart of nSure.ai’s business model. Rather than merely providing fraud prevention software, the company assumes financial liability for the transactions it approves. If nSure.ai recommends that a transaction should be approved and that transaction later proves fraudulent, the company reimburses the loss.
That changes the relationship between fraud prevention providers and merchants entirely. Most fraud companies are paid regardless of the outcome. nSure.ai effectively puts its own balance sheet behind its decisions. The better the merchant performs, the better nSure.ai performs. As Podvin describes it, the interests of both parties are completely aligned.
The idea emerged from experience rather than theory. Before launching nSure.ai, Podvin and his colleagues operated a digital gift card business. The venture appeared successful until they discovered that a significant proportion of their revenue had disappeared through fraud. Existing solutions failed to solve the problem, so they built their own system.
“We built something very pragmatically to solve our own problem,” he says.
What emerged differed significantly from traditional fraud prevention models. Historically, the industry focused on authenticating digital identities. The objective was straightforward: confirm that the person conducting a transaction was genuinely who they claimed to be. Today, Podvin argues, that approach is increasingly ineffective.
“What we see in crypto is that 80% of fraud comes from verified accounts.”
It is a striking statistic because it challenges one of the core assumptions of modern compliance. Companies spend enormous resources verifying customer identities, yet many of the most sophisticated fraudsters have already passed those checks. They possess legitimate documentation, verified accounts and increasingly sophisticated technology.
As a result, nSure.ai spends less time asking who somebody is and more time examining what they are doing.
nSure.ai looks at the world through hundreds of thousands of compounded data points. Podvin uses the analogy of a spider’s web. Imagine a web so fine that any movement anywhere within it can be perceived and analyzed. Individual transactions may appear entirely legitimate when viewed in isolation, but patterns begin to emerge when thousands of transactions are analysed collectively. It is those patterns, rather than individual events, that often reveal fraudulent behaviour.
The challenge is becoming increasingly complex as artificial intelligence transforms the fraud landscape. Social engineering, where victims are manipulated into willingly transferring money to criminals, has existed for decades. What has changed is scale.
Previously, a fraudster might make fifty calls in a day attempting to persuade victims to part with their money. Today, AI-driven systems can make millions.
“One person could call fifty people. Now the machine can call five million, with equivalent success rates.”
That shift has profound implications for the financial industry. Social engineering is no longer a niche criminal activity. It has become industrialised.
The most immediate challenge lies in Authorised Push Payment fraud, or APP fraud, where victims are persuaded to transfer money directly into a fraudster’s account. Unlike traditional card fraud, there is often no intermediary standing between the sender and recipient. The transaction is authorised, the payee is perfectly legitimate, and the payment is processed exactly as instructed. The victim only discovers later that they have been manipulated.
For years, this created a regulatory grey area. If the customer willingly initiated the transfer, who was responsible for the loss? According to Podvin, the answer was frequently nobody beyond the victim themselves.
That is beginning to change. Regulators are increasingly moving towards frameworks that assign liability to the financial institutions at either end of the transaction. Once banks become financially responsible for APP fraud losses, the economics of fraud prevention change dramatically. For years, APP fraud existed in a liability vacuum where the victim often carried the loss. By assigning responsibility to financial institutions, regulators are creating a powerful commercial incentive to invest in prevention. What was once a customer problem becomes a banking problem, and ultimately the potential to apply solutions to identify fraud before the money moves.
At the same time, two other massive transformations are underway:
For one, agentic commerce (AI agents transacting on behalf of human users) is becoming the next payments frontier, generating important questions including “is this agent authorized by a real human?”, “does the human want to perform the transaction the agent is planning to execute?”, “is this merchant legitimate in an agent-to-merchant context?”
For nSure.ai that today processes many transactions executed by bots, this is not a problem. The authentication question might be solved by the emerging protocols such as X402, AP2, or MPP, designed precisely to do that. But what matters is the intent, a question that nSure.ai excels at, with superior accuracy.
The other major change is that the worlds of traditional banking and cryptocurrency, once viewed as ideological opposites, are beginning to converge. Crypto exchanges are acquiring banking licences, while established financial institutions are embracing digital assets, stablecoin, and blockchain infrastructure.
“The two worlds are merging into what is now the new financial services,” says Podvin.
For companies operating at the forefront of fraud prevention, that convergence is significant. The challenges that first appeared in crypto are increasingly becoming mainstream banking challenges. The observation deck that crypto provides into emerging fraud trends has never been more valuable.
None of this means the battle is won. Podvin readily acknowledges that fraud prevention remains an arms race. Fraudsters are adopting artificial intelligence, automation and increasingly sophisticated behavioural techniques. Defenders are responding with advanced analytics, machine learning and new approaches to risk management.
Yet he remains optimistic. The future, he believes, lies not simply in identifying fraud but in fundamentally changing its economics. By aligning incentives, assuming liability and focusing on behaviour rather than identity alone, companies such as nSure.ai.ai are attempting to move the industry beyond traditional fraud prevention and towards something more ambitious.
The fraudsters are evolving and the defenders have little choice but to evolve faster.
Pascal Podvin will be speaking at the House of Block digital assets conference in Ham. He has denied rumours that he is bringing his own polo pony.
