As AI agents increasingly influence online shopping, new threats are emerging for e-commerce merchants. To address these risks, Riskified, a global fraud prevention company, and Human Security, a digital experience protection provider, have partnered. They introduce a unified framework designed to safeguard merchants from AI-driven fraud, revenue loss, and reputational damage.
The framework combines Human Security’s trust layer—enabling visibility and governance of AI shopping agent interactions—with Riskified’s expertise. It covers fraud prevention, chargeback protection, and policy abuse. According to the companies, this collaboration allows merchants to approve more legitimate AI-driven orders. It also helps reduce false declines and protect profit margins.
Key risks highlighted include:
- Fraud evasion: AI agents can bypass traditional fraud filters by mimicking human behavior and adapting to static rules.
- LLM-referred traffic: Riskified’s data shows large language model (LLM) traffic can be up to 2.3 times riskier than standard search traffic.
- Reseller arbitrage: Bots can strip inventory at scale, reselling goods like sneakers or tickets at inflated prices, disrupting pricing strategies.
- Trust erosion: Over-reliance on bots risks undermining consumer confidence in site security and authenticity.
Industry experts warn that rules-based fraud detection is increasingly inadequate, as AI agents can probe thresholds, solve captchas, and dynamically shift tactics. Analysts also caution that merchants must balance fraud defenses with user experience, as heavy-handed controls may further erode customer trust.
Despite these challenges, the collaboration between Riskified and Human Security is seen as a step toward creating layered defenses that combine AI-driven analytics, anomaly detection, and governance frameworks. Experts agree that maintaining customer trust will be the ultimate differentiator as AI reshapes e-commerce.
Source:
