The next wave of ecommerce innovation has arrived, ushered in by the rapid rise of agentic artificial intelligence. Powered by generative AI, agentic shopping — where digital agents transact with online stores on behalf of customers — is creating unprecedented opportunities for growth. Yet, for all the promise it holds, this new segment of commerce also introduces significant risks, particularly in the battle against fraud and abuse.

Riskified is at the forefront of this transformation, launching multiple tools designed to help ecommerce merchants trust good customers with legitimate AI shopping agents while protecting their businesses from malicious automation from bots, scrapers, and fraudsters using compromised credentials. 

Through strategic partnerships, innovative technology, and enhanced infrastructure, Riskified is empowering merchants to adapt securely and efficiently to the future of agentic commerce.

At the core of these advancements are four key components:

  1. HUMAN Security Partnership: A shared vision to advance a unified security framework that helps merchants win and grow safely via emerging agentic channels. 
  2. AI Agent Approve: The MCP Server Package on AWS Marketplace allows supply-side (merchants) and demand-side (LLMs) parties to communicate with the Riskified platform APIs and accelerate the safe adoption of AI shopping agents.
  3. AI Agent Intelligence: Dashboard views in the Riskified Control Center for monitoring ecommerce orders originating from AI shopping agents.
  4. AI Agent Policy Builder: Configuration and enforcement of agentic policy abuse in the Riskified Decision Studio, including building, simulating, deploying, and tracking multiple types of business logic and policies use cases.

These three solutions interweave a security infrastructure built on the foundation of Riskified’s expertise in fraud prevention and ecommerce enablement, delivering precise decisions and improved business performance for merchants.

01. Partnership announcement with HUMAN Security

Riskified announced a joint commitment with HUMAN Security to leverage their respective industry-leading AI platforms and expansive network insights to secure the next era of digital commerce.  

By aligning HUMAN’s recently launched HUMAN Sightline featuring AgenticTrust with Riskified’s ecommerce risk management expertise in fraud prevention, chargeback protection, and policy abuse prevention, merchants can apply consistent trust policies and transaction decisions across both human and AI-driven interactions. This enables them to confidently accept more legitimate AI-driven purchases, block sophisticated fraud, and protect customer relationships.

“The ‘do it for me’ AI agent revolution is happening, but the shape of agentic commerce is still emerging,” said Assaf Feldman, CTO and Co-Founder of Riskified. “In a world where AI agents transact on behalf of individuals, resolving identity and trust becomes more complex. By working with HUMAN and developing new agentic tools and capabilities, we give merchants a way to safely embrace this shift, turning what could be a threat into a new, profitable digital channel.”

Partnership announcement with HUMAN Security
*Simplified architecture for visual clarity

02. AI Agent Approve, a Model Context Protocol (MCP) Server

With OpenAI’s July 2025 announcement that they intend to launch their own browser, and Perplexity AI’s recent launch of Comet, there is an evolving set of challengers to Google Chrome’s dominance, with the potential to reimagine the browser experience. For example, instead of users searching Google and immediately clicking through to other websites, some user interactions and context may remain within a ChatGPT-like native chat interface.

If and when OpenAI adds secure payment wallet capabilities to its browser, the foundation will be set to unleash the “do it for me” AI shopping revolution, allowing agents to select goods and securely purchase them without human intervention.

However, with this exciting innovation comes new threats. For example, compromised LLM accounts from account takeovers (ATOs), stolen payment methods loaded into legitimate AI agent accounts, AI agents used by bad actors to perform sophisticated promotion abuse, or reseller abuse at scale.

AI Agent Approve provides a trust layer written in the open standard MCP to help the demand-side AI shopping agent establish its credentials in a language it understands. For example, suppose you’re a demand-side AI agent, from an LLM like OpenAI’s ChatGPT. In that case, you can communicate directly with Riskified’s AI Agent Approve MCP to get a fraud decision leveraging the Riskified risk intelligence platform and merchant network.

For supply-side (merchant) agents, the merchant’s agent can call the AI Agent Approve MCP and request a decision on promotion abuse based on the identity we see from that consumer-side agent. As this technology evolves and further security protections mature, there will be opportunities to adapt our MCP to provide selective verification when certain risk factors indicate that a bad actor may have taken over a consumer’s shopping agent.

03. AI Agent Intelligence in the Riskified Control Center

The surge in AI traffic is leaving many merchants struggling to differentiate between helpful agents and harmful bots, creating blind spots in understanding emerging customer trends and linking AI-driven traffic to outcomes like payment fraud, abuse, or policy violations.

Riskified’s new AI Agent Intelligence in the Riskified Control Center bridges this gap. For the first time, merchants gain visibility and control into AI traffic, identifying which LLM platforms drive real revenue versus risk. Powered by our proprietary models, behavioral analytics, and our partnership with HUMAN, Riskified’s platform goes beyond basic monitoring to help merchants embrace legitimate AI shoppers while blocking sophisticated threats.

04. AI Agent Policy Builder in the Riskified Decision Studio

Riskified’s AI Agent Policy Builder equips businesses to stay ahead of emerging threats in an AI-driven landscape. Built into the Decision Studio, this capability helps to address evolving abuse patterns driven by AI agents before they impact business performance. AI agents could drive higher returns, scale promo abuse, and make life easier for resellers looking for items on sale. Identifying early signals of new behaviours and forms of policy abuse driven by AI allows merchants to address root causes, not just symptoms.

Merchants gain the flexibility to fine-tune site policies with precision, protect margins, and optimize platforms specifically for AI-layer interactions. By monitoring agent behavior, businesses can distinguish benign AI-driven traffic from agents exploiting vulnerabilities.

With this depth of analysis, merchants can build, simulate, track, and adjust custom controls for any AI source or pattern. Every change is measurable, empowering confident decisions before deploying updates. The result is a proactive approach, transitioning from reacting to abuse to actively setting the rules of engagement, ensuring policy enforcement, profitability, and resilience in today’s dynamic ecommerce environment.

Ready to embrace the future of ecommerce with confidence?

Contact us today to learn how Riskified’s solutions can help you innovate securely and thrive in the era of AI shopping agents.