New study finds business email compromise losses average $224K
Business email compromise tied to artificial intelligence produced the highest average losses per complaint among AI-related crimes in 2025. Simmrin Law Group analysis of FBI Internet Crime Complaint Center data placed the figure at $224,123 for each of the 135 incidents, 54.5 percent above the $145,097 average recorded for investment fraud. The category accounted for just 0.6 percent of the 22,364 total AI-related complaints yet generated $30.26 million in reported losses.
The underlying table showed BEC complaints produced $30,256,592 in losses and represented 3.4 percent of all AI-related losses. Investment fraud generated $632,041,188 across 4,356 complaints and accounted for 70.7 percent of the $893,346,472 national total. The overall average loss per complaint across all AI-related categories stood at $39,946.
The contrast between complaint volume and financial impact is pronounced for BEC. Although the 135 complaints formed only 0.6 percent of the AI-related total, they produced 3.4 percent of losses. Categories with far higher complaint counts, such as personal data breach with 1,204 cases, recorded lower per-complaint losses of $15,588. Government impersonation generated an average of $27,160 across 260 complaints.
BEC losses per complaint exceeded those of several other categories by wide margins. The $224,123 average reached 6.6 times the $33,897 recorded for tech and customer support scams. It was nearly ten times the $23,470 average for real estate fraud and more than twelve times the $18,162 average for employment scams. Phishing and spoofing complaints averaged $12,807, making BEC losses 17.5 times higher.
Broader FBI statistics provide additional context for the AI-related subset. Across all BEC methods in 2025, authorities recorded 24,768 complaints carrying $3.05 billion in losses. The 2024 totals were 21,442 complaints and $2.77 billion in losses. Reported losses rose about 10 percent while complaint volume increased 15.5 percent. These figures cover both traditional and AI-assisted schemes.
The FBI describes BEC as schemes in which criminals manipulate legitimate business communications to obtain unauthorized payments. Methods include compromised accounts, impersonated executives, altered vendor instructions, and lookalike email addresses. AI-generated language can increase the apparent legitimacy of such messages. The 2025 data does not identify which individual cases used these techniques. The agency advises businesses to verify payment instruction changes through independently confirmed contact information rather than replying to the requesting email.
The reported averages carry important limitations. They reflect mean losses across the complaints in each category and do not represent typical victim outcomes. A small number of large incidents can raise the average substantially. The FBI data also does not specify the AI techniques used in individual cases, preventing determination of how artificial intelligence affected outcomes in the BEC category.
Simmrin Law Group performed the analysis by examining the FBI’s 2025 Internet Crime Complaint Center dataset. The review compared complaint counts, reported losses, shares of total AI-related losses, and losses per complaint across the ten categories to distinguish where financial harm concentrates from how frequently complaints occur.
