E-commerce is entering a new phase of fraud, where transactions that appear legitimate — valid accounts, correct details, even routine behavior — can mask coordinated attacks powered by synthetic identities and AI.
From account takeovers (ATOs) to loyalty abuse, fraud now targets identity rather than the point of sale, warned Dany Naigeboren, senior director of risk at fraud prevention company Forter.
Fraud is no longer breaking into systems — it’s operating inside them.
These attacks are becoming easier to execute, allowing smaller actors to operate with capabilities once limited to organized fraud rings.
“It’s a huge game changer. Until now, online fraud has mostly involved sophisticated fraud syndicates with vast engineering resources,” Naigeboren told the E-Commerce Times.
To counter that threat, Forter uses AI to detect and prevent “invisible” fraud across the full customer journey, not just at checkout.
Its global merchant network draws on data from over $1 trillion in transactions to improve detection accuracy, reflecting a broader need for AI-driven systems to track and stop increasingly difficult-to-detect attacks.
Digital Thieves Use AI Against Us
Table of Contents
AI is lowering the barrier to entry for e-commerce fraud, allowing individuals with limited technical skills to execute attacks once limited to organized teams.
“With two to three hours of work, anyone can become proficient in conducting fraud online. And this manifests itself in numerous ways,” Naigeboren said.
In many cases, the fraud no longer looks like an obvious attack. Compromised accounts, valid credentials, and routine purchasing behavior allow transactions to move through ordering and payment systems without triggering traditional defenses.
Attackers are increasingly bypassing merchant systems by targeting consumers directly, using account takeovers and similar schemes to operate inside legitimate user activity.
AI tools also make it easier to carry out social engineering and phishing attacks, helping fraudsters steal credentials and impersonate users at scale.
“We’re suddenly getting more emails targeting us to get information to conduct fraud online. This is one area that has changed immensely,” he said.
He added that AI is making fraud both easier to execute and more profitable, enabling attackers to deploy automated agents that scale these activities rapidly.
Fighting Fire With Fire
Forter’s approach centers on identity — determining who is behind each interaction and whether that intent is legitimate.
“The way we see it is that to conduct fraud using sophisticated AI tools, you need equally advanced systems to stop it,” Naigeboren said.
While the platform is primarily used in e-commerce, it also supports hybrid models such as buy online, pick up in store, where online identity signals extend into physical transactions.
It provides real-time decisions that approve or decline transactions, reducing false declines while addressing account takeovers, fake accounts, payment fraud, and promotional abuse.
AI Redirects Fraud Toward Consumers
Naigeboren is seeing spikes in account takeovers and social engineering, with phishing and impersonation attacks growing more convincing and harder to detect.
Rather than targeting merchant systems directly, many fraudsters now focus on consumers, using compromised accounts to move through e-commerce workflows with fewer controls.
In practice, that can mean a legitimate account placing what appears to be a routine order — but one initiated by an attacker using stolen credentials.
“It’s often easier and more lucrative to target consumers directly,” Naigeboren said.
AI also enables attackers to deploy automated agents that scale these tactics, reducing the need for technical expertise while increasing attack volume.
At the same time, retailers face rising “friendly fraud,” where legitimate-looking customers abuse policies, promotions, and chargebacks.
The challenge, Naigeboren noted, is determining intent — whether a user is making a legitimate purchase or attempting to game the system.
Forter addresses this in real time and through post-transaction analysis of returns and chargebacks, which remain among the most difficult areas to manage.
How Forter Differs From Competitors
Effective fraud detection depends on combining real-time signals with historical context to understand user behavior.
Most systems evaluate transactions in isolation. Forter instead connects behavior across interactions, linking activity across sessions, devices, and transactions to build a unified view of each user.
That allows the platform to assess risk at the identity level, instead of relying on individual transactions or single touchpoints.
“Traditional fraud detection systems are siloed. They look at one or two data points at a single touchpoint. Modern fraud prevention needs a holistic, identity-centric view over multiple touchpoints such as checkout, returns, promotions, and logins,” he said.
The Future of Fraud Prevention
E-commerce fraud is expanding rapidly, with new attack methods emerging faster than traditional defenses can adapt.
“There is no single silver bullet. The key is holistic visibility into interactions and identities. Detection and protection systems must be AI-driven and continuously learning, not static rule sets,” Naigeboren said.
As AI agents increasingly transact on behalf of users, the central challenge is identifying who — or what — is behind each interaction.
In some cases, those interactions may look indistinguishable from legitimate customer behavior, with automated agents browsing, adding items to carts, and completing purchases without obvious signs of fraud.
Distinguishing between human users and AI-driven activity remains inconsistent across systems, often relying on fragmented signals or indirect detection methods.
“As AI agents handle more checkouts and logins, it becomes essential to tie each agent’s activity back to a verifiable underlying human identity,” he said — a requirement that is quickly becoming central to modern fraud prevention.
