Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Darwinium Launches AI Agent Fraud Detection Technology – Digital Transactions in Simple Termsand what it means for users..
Darwinium Holdings Inc., an artificial intelligence-based digital security and fraud-prevention platform, launched its AI agent fraud-detection solution early Tuesday. The technology is expected to help businesses combat the growing threat from AI-facilitated fraud as agentic commerce grows in popularity.
Research conducted by Darwinium reveals that 97% of respondents reported an increase in AI-based fraud attacks in the past year. In addition, respondents estimated that 40% of the fraud attempts they see today use artificial intelligence. Darwinium surveyed 500 fraud and risk leaders.
“Agentic commerce is already here…and we are already seeing a meaningful portion of [our customers’] traffic coming from AI agents,” Alisdair Faulkner, co-founder and chief executive of Darwinium, says by email. “At the same time, this growth is being mirrored by a rise in AI-driven fraud. Clearly, businesses require solutions that can distinguish between agents acting legitimately on behalf of customers and those acting with malicious intent.”
As AI-driven fraud becomes more common, traditional fraud-prevention defenses against software-driven fraud—such as blocking all non-human traffic entirely from purchasing on a site—is proving ineffective and costly for businesses.
“Blocking all non-human traffic is no longer practical because many consumers are using AI agents for legitimate purposes,” Faulkner says. “In that environment, the bot strategy risks blocking your best customers’ automated assistants along with malicious actors.”
Such a strategy can cost businesses millions of dollars in lost sales annually, Faulkner says.
Darwinium’s AI fraud-detection technology seeks to authenticate the AI agent and evaluates its behavioral intent prior to purchase using five layers of intelligence: device signatures, network telemetry, timing patterns, behavioral biometrics, and journey analysis. Businesses can define fraud detection and prevention to allow specific trusted agents to proceed, set thresholds, and require human confirmation for sensitive actions.
“By combining these signals, we create confidence-based assessments of activity rather than relying on a single binary indicator,” Faulkner says. “This multi-layered approach allows us to distinguish trusted automation from adversarial agent behavior across the entire customer journey.”
Darwinium developed its AI agent fraud-detection technology to be interoperable with evolving payment and identity protocols that support agentic commerce, including Visa Inc.’s Trusted Agent Protocol, Mastercard Inc.’s Agent Pay, and emerging know-your-agent models.
“Understanding user intent across the full customer journey provides roughly 30%–40% greater visibility into user and agent behavior compared to traditional tools that protect only isolated touchpoints,” Faulkner says. “A more contextual, journey-based approach helps reduce fraud while preserving legitimate customer activity.”