Skyfire and Forter are expanding their partnership to help online retailers distinguish legitimate AI shopping agents from malicious automated traffic during checkout. The integration adds Skyfire’s Know Your Agent (KYA) identity data to Forter’s fraud decisioning, aiming to reduce false declines of authorised agent-driven purchases without increasing exposure to fraud.
The companies say the integration was piloted with Jomashop, an online retailer of luxury watches, jewellery and accessories. During the pilot, Jomashop achieved a nearly 100% success rate for agentic checkouts, according to the announcement, without an increase in chargebacks. The companies have not disclosed the pilot’s transaction volume, measurement period or independent verification of the result.
How the Integration Works
Traditional bot-management systems often treat automated browsing as a potential threat. That approach becomes more complicated when a customer delegates shopping to an AI agent that searches products, selects an item and initiates a purchase on the customer’s behalf.
Skyfire’s Know Your Agent protocol is designed to provide identity information about an AI agent and the person or organisation it represents. Forter incorporates those signals into its existing fraud decisioning, alongside other transaction and customer information. The aim is to help merchants distinguish authorised agent activity from suspicious automation rather than blocking traffic solely because it is non-human.
According to the companies, the new signals pass alongside existing risk data, so merchants do not need to introduce a different checkout interface. The partnership builds on Forter’s Fraud Management for Agentic Commerce capabilities, which the company announced earlier in October.
The timing reflects a shift in online shopping behaviour. Forter reported that agentic orders across its network had increased more than fourfold since August, while automated fraud had doubled over the preceding two months. These figures come from Forter’s own network and have not been independently audited.
Jomashop Pilot Reports Near-Complete Checkout Success
The companies say the integration was piloted with Jomashop, an online retailer specialising in luxury watches, jewellery and accessories. Before the integration, conventional bot-management controls reportedly blocked much of the agent-driven traffic.
Following the implementation, Jomashop achieved a nearly 100% checkout success rate for the relevant agentic traffic without an increase in chargebacks, according to the announcement. The result suggests that better agent identification may help retailers recover legitimate transactions that generic automation controls would otherwise reject.
However, the companies have not published the number of transactions in the pilot, the baseline success rate, the duration of testing or the precise definition of checkout success. The result therefore should not be interpreted as a guarantee of near-perfect conversion or fraud prevention across other retailers.
For merchants, the relevant question is whether identity signals can improve approval rates while keeping fraudulent purchases and chargebacks within acceptable limits. That requires ongoing monitoring as agents, fraud techniques and transaction patterns evolve.
Trust Becomes Infrastructure for Agentic Commerce
The integration reflects a broader industry effort to make agent-driven shopping compatible with existing commerce and security infrastructure. Skyfire has also worked with companies such as F5 and Fastly to help businesses identify trusted agent traffic without automatically treating every automated interaction as malicious.
That work points to a developing requirement for agentic commerce: retailers need evidence of which agent is acting, who authorised it and whether the transaction fits the underlying customer’s behaviour. Identity alone cannot establish that a purchase is safe, so fraud models must combine agent credentials with transaction-level risk signals and merchant-specific policies.
For retailers, the commercial opportunity is to accept legitimate purchases initiated by AI systems without weakening existing protections. For the broader ecosystem, the Skyfire-Forter partnership offers one approach to integrating agent identity into fraud decisions.
Its longer-term value will depend on independent performance data, adoption across agent platforms and whether the same safeguards work at scale. As AI assistants become more involved in purchasing, trusted identity may become an important part of checkout infrastructure—but it will need to work alongside, not replace, established fraud controls.
Market Landscape
Agentic commerce is creating a new challenge for e-commerce security: legitimate AI agents can resemble the automated traffic that retailers have historically blocked. Identity protocols, agent-aware fraud models and new checkout interfaces are emerging as complementary ways to address that problem.
Skyfire supplies agent identity signals, while Forter applies those signals within its fraud decisioning platform. Their integration illustrates how established fraud systems may adapt to purchases initiated by software rather than directly by a human shopper.
Top Insights
- Agent identity: Skyfire’s KYA data helps Forter assess who an agent represents when it initiates a purchase.
- Retailer pilot: Jomashop reported nearly 100% success for the relevant agentic checkout traffic, according to the companies.
- Existing checkout preserved: The integration passes agentic signals alongside existing risk data without changing the checkout interface.
- Fraud remains a concern: Forter reports that automated fraud has doubled over two months, highlighting the risks of accepting agent traffic indiscriminately.
- Evidence is limited: The companies have not disclosed pilot transaction volumes, detailed methodology or independent validation of the reported results.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
