Guarding the Lifecycle of Agentic Payments and Multi-Layered Commerce

At the FTT Payments 2026 event, the discussion surrounding transaction risk highlighted a major shift in how commerce is executed and secured. Tom Taraniuk, Head of Partnerships at Sumsub, detailed how the convergence of advanced artificial intelligence and automated consumer behavior is fundamentally reshaping the global payments landscape. While the industry frequently cycles through various technology buzzwords, the rapid rise of AI is driving a sophisticated evolution in fraud patterns while simultaneously giving birth to a massive new transactional framework: agentic commerce.

The Democratization of Fraud and AI-Driven Exploits

Artificial intelligence has fundamentally altered the threat matrix for financial institutions by democratizing access to high-level disruption tools. Historically, executing highly technical cyberattacks or constructing believable identity bypasses required specialized engineering skills. Today, AI has commoditized these capabilities, placing powerful deepfake software, generative code, and “fraud-as-a-service” toolkits directly into the hands of opportunistic fraudsters.

This democratization serves as a massive game-changer for casual opportunists and highly organized criminal syndicates alike. Nefarious actors actively deploy automated scripts and generative tools to target fintech onboarding flows, compromise legitimate user funds, and systematically inject “sleeper accounts” into banking systems. These sleeper profiles are designed to pass initial, static Know Your Customer (KYC) checks, sitting dormant until criminal networks take them over to route illicit funds and circumvent traditional anti-money laundering (AML) rules.

The Rise of Agentic Payments and High-Value Delegated Commerce

While AI presents severe security challenges, it is also powering the single largest transactional paradigm shift of the decade: agentic payments. Global payment networks, including Visa and Mastercard, project that autonomous AI agents will command $20\%$ of the entire payments market and transaction flow by 2029.

“This is something which is not going to just be your daily toilet paper order from Amazon. You’ll be using agents to go and book trips at merchants or travel agencies up to 5,000 euros.”

This projection underscores that delegated commerce is moving far beyond low-value, repetitive grocery or household refills. Consumers will routinely deploy autonomous AI agents as financial proxies, tasking them with navigating travel agencies, selecting merchants, and executing complex, high-value transactions up to €5,000.

This evolution introduces critical operational hurdles for modern payment processors. When a machine transacts on behalf of a human, traditional authentication cards and biometric checks fail. The financial ecosystem urgently requires new verification protocols to establish the identity of the digital agent, validate the true intent of the human owner, and firmly mitigate the risk of post-transaction chargebacks. Major networks like Visa and Mastercard are actively examining these exact protocol requirements to safely accommodate automated machine-to-machine spending rails.

Protecting the User Lifecycle with Multi-Layered Intelligence

The primary risk of the agentic era stems from the fact that autonomous agents are highly intelligent software engines. When backed by legitimate consumers, they drive unprecedented economic efficiency; however, when weaponized by bad actors and criminal networks, they can scale business exploits at a velocity that traditional human compliance teams cannot match. Static, point-in-time security checks executed only during the initial signup phase are completely obsolete against an automated, adaptive threat matrix.

To enable the industry to safely adopt agentic commerce, Sumsub is helping financial institutions deploy comprehensive, multi-layered identity and fraud protection frameworks. True defense requires monitoring the entire end-to-end lifecycle of the customer journey, combining upfront verification with continuous behavioral analysis thereafter.

By layering real-time device signals, biometric validation, and ongoing transaction monitoring, the platform flags anomalous agent activity and sleeper account takeovers the moment they deviate from normal human parameters. This multi-layered stance allows banks and fintechs to successfully neutralize AI-driven fraud and secure automated transaction flows, ensuring the future of agentic payments remains safe, compliant, and fraud-resistant.

Key Highlights from Tom Taraniuk:

• The Fraud-as-a-Service Boom: Taraniuk warns that AI has democratized access to advanced deepfake and circumvention tools, drastically lowering the barrier to entry for opportunistic fraudsters.

• The $20\%$ Market Projection: Major payment networks like Visa and Mastercard forecast that autonomous AI agents will process one-fifth of all global payment flows by 2029.

• High-Value Machine Transactions: Moving past low-value retail orders, digital agents will independently authorize high-complexity purchases, such as travel bookings up to €5,000.

• The Sleeper Account Menace: How organized criminal networks exploit automated onboarding flows to plant dormant accounts, evading traditional banking rules for later exploitation.

• The Need for Agent Protocols: Highlighting the urgent industry mandate for card networks to establish specialized verification and intent protocols to block machine-driven chargebacks.

• Continuous Behavioral Defense: Why financial institutions must move away from static onboarding checks and implement multi-layered, lifecycle-wide monitoring to defeat smart AI-driven attacks.