AI-powered fraud hitting companies bottom lines, finds KPMG survey
AI-powered fraud is emerging as serious threat to Canadian organizations, according to a recent KPMG Canada survey, with 72% of companies losing as much as 5% of annual profits to AI scams. The Big Four firm surveyed 251 Canadian companies in February 2026.
As the fraud landscape takes on additional complexity, 81% of companies that experienced fraud in the last year say they faced an AI-enabled attack.
Nearly all business leaders (94%) are concerned about encountering AI-enabled attacks in the coming year, but only a quarter (26%) have a tested response plan to defend against AI-enabled attacks such as deepfakes and voice clones.
“Canadian organizations aren’t just seeing more attempted attacks – they’re more sophisticated, harder to spot and faster to execute, leaving many businesses vulnerable and unprepared to fight back,” said Myriam Duguay, partner and national leader of forensic investigation, integrity and dispute services at KPMG Canada. “Now, with rise in AI-powered attacks that can mimic legitimate business interactions with alarming accuracy, the margin for error becomes razor-thin and having strong fraud defences is even more essential.”
The most common attacks respondents faced were AI-generated phishing emails/chats (60%), deepfake documents (39%), and voice-clone executive impersonation calls (24%).
With AI attacks becoming more frequent and hard to detect, companies are using AI as a defensive tool, with approximately half (52%) using AI to “fight back” by using it to identify anomalies, authenticate users, and detect manipulated content.
Six in 10 respondents plan to increase fraud detection and prevention budgets by up to 7%, while eight in 10 say they conduct employee fraud awareness training every six-to-12 months.
“While momentum is building, many organizations still have significant work ahead to fully modernize and strengthen their defences against rapidly evolving AI-enabled fraud,” says Marilyn Abate, a partner in KPMG Canada’s risk services practice.
Duguay and Abate say that, in addition to staff training, there are several key areas companies can address to improve anti-fraud programs in the AI age.
First, companies should think of fraud as a strategic capability that includes governance, talent, and accountability. Next, companies should avoid relying on point-in-time checks and should instead use continuous, risk-based controls to prevent and detect fraud earlier in the lifecycle.
