Contents
Key fraud trends shaping ecommerce in 2026
A collection of shareable ecommerce fraud statistics based on key findings from Signifyd’s State of Fraud 2026 report.
North America Signifyd data, Jan. – April 2026 vs. Jan. – April 2025
AI-powered fraud trends are flourishing
For the first time in the 26-year history of its annual cybercrime report, the FBI recently included a section on AI-generated scams, which cost consumers nearly $900 million in 2025.
Just as AI is reshaping how consumers research, discover and buy online, fraudsters are exploiting the same technology to move faster, scale attacks and increase the profitability of their schemes. In fact, Interpol estimates that AI-powered fraud schemes are 4.5 times more lucrative than conventional attacks.
AI-powered fraud schemes are 4.5x more lucrative than conventional attacks
Source: Interpol
What used to require teams of specialists working through slow, manual trial and error can now be quickly executed with inexpensive, sophisticated AI tools. And data from Signifyd's Commerce Network shows the move from manual fraud operations to AI-assisted automation is having a clear impact: Fraud pressure in North America alone increased 33% year over year in the first four months of 2026. First-party fraud and abuse also grew 9% during the same period.
Account takeover attempts also surged 78% between Jan. – April 2026 — an evident sign that automated credential testing at scale is now the norm.
+78% account takeover attempts
(Jan. – April 2026 vs. Jan. – April 2025)
Source: Signifyd data
On top of that, post-purchase abuse is getting easier. Generative AI (gen AI) can now create authentic-looking receipts complete with store branding and generate hundreds of images of shattered, damaged or defective products on demand, making false return and refund claims easier to execute. In fact, “item not as described” claims rose 49% in the first four months of 2026 vs. the same period in 2025, according to Signifyd data.
+49% increase in "item not as described" claims
(Jan. – April 2026 vs. Jan. – April 2025)
Source: Signifyd data
Humans are the newest fraud attack vector
Automated social engineering has made people the latest attack surface. The tools allow each link in the fraud supply chain to work at a scale and speed never before possible.
“AI-enabled synthetic content is becoming increasingly difficult to detect.”
Con artists who rely on phishing, smishing, vishing and quishing can acquire personal information faster than ever. And deepfakes using gen AI make these attacks even more convincing. Criminals can impersonate victims' loved ones in video calls or voice messages, creating scenarios that feel authentic enough to trick people into handing over credentials or financial information.
The volume of attacks reflects the ease of execution. Nearly 200,000 people fell victim to phishing scams last year, costing them $215.8 million, according to the FBI's latest cybercrime report.
Fraud is now a connected supply chain
In 2026, each fraud scheme is part of an integrated criminal ecosystem, and democratized AI has wired the pieces together into a single supply chain of misery.
Before fraudsters can use their haul, they need to know what's valid. Lightning-fast, AI-powered carding attacks now certify in moments whether stolen credentials are ready for fraudulent purchases. In fact, data from Signifyd's Commerce Network shows that AI is accelerating the validation process at a massive scale: Card-testing attacks rose 175% year-over-year between Jan. – April in 2026.
Card-testing attack increase
(Jan. – April 2026 vs. Jan. – April 2025)
Source: Signifyd data
Once validated, the stolen information fuels account takeover fraud, synthetic identities and fraudulent purchases. With account takeover, fraudsters break into aged accounts, shielded by trustworthy account history, and purchase goods with stored payment methods or drain loyalty points. Deepfakes help put convincing faces and voices to synthetic identities, making them harder to detect. Professional-grade fake websites spring up to sell stolen goods and harvest the next round of credentials. Then the cycle starts again.
BOPIS fraud is picking up
BOPIS (buy online, pick up in store) schemes involving mules are flourishing. Signifyd data found that BOPIS fraud rose 65% in the first third of 2026.
BOPIS fraud increase
(Jan. – April 2026 vs. Jan. – April 2025)
Source: Signifyd data
Typically, in these schemes, online orders for large volumes of little-scrutinized items are picked up and later returned for store credit. The freshly laundered store credit is used to avoid fraud review when it's applied to purchase highly fenceable items — think a houseful of smoke detectors exchanged for expensive, easily resellable power tools.
2026 fraud trends by region
Europe and the UK
Merchants in Europe and the UK are living the reality of shifting online fraud threats. In the first four months of 2026, account takeover attempts increased by 67% year-over-year, according to Signifyd data, and post-purchase abuse climbed. Claims that an item never arrived rose 49%, while claims that a delivered product wasn't as described increased by 35% in the same period.
Account takeover
AI increases speed and scale
Item not received
Researches vulnerabilities
Item not as described
AI doctors photos
Jan. – April 2026 vs. Jan. – April 2025
Source: Signifyd data
Latin America
Latin America is a case study in how AI has industrialized fraud. Over the past year, organized schemes grew steadily more sophisticated, shifting the battlefield from untargeted attacks to optimized, higher-value heists.
Fraud average order values (AOV) increased 3% during the first four months of 2026, according to Signifyd data, compared to the previous year. LATAM fraud pressure also rose 7.5% during the same timeframe.
Consumer abuse — false claims an order never arrived or the product was not as described, for instance — skyrocketed 168% during the same period.
Increase in consumer abuse
(Jan. – April 2026 vs. Jan. – April 2025)
Source: Signifyd data
Country-by-country and vertical-by-vertical, the playbook for AI-powered fraud schemes varied dramatically. Mexico saw a rise in ATO targeting mature accounts while Chile experienced more rapid-fire, lower-AOV attacks. Crime rings in Brazil targeted same-day delivery platforms for card-testing attacks, while Argentina saw the region’s highest increase in low-value attack frequency.
STATE OF FRAUD 2026
Read the State of Fraud 2026 report to get ahead of today’s ecommerce fraud trends
Download the full report for the complete data, industry trends, regional breakdowns and true-crime case studies.
