"My AI Bought That, Not Me" — Chargebacks911 Warns of New Dispute Wave in Agentic Commerce Era

Akihiro Suzuki

Akihiro Suzuki

Twitter

Key Takeaways

  1. Chargebacks911 warns of new chargeback categories driven by AI agent-initiated transactions
  2. Dispute resolution frameworks lag behind as Visa and Mastercard advance agentic payment pilots
  3. E-commerce businesses must urgently build agent permission controls, transaction audit trails, and notification systems

AI Agents Now Execute Purchases, Fundamentally Changing Dispute Patterns

'I didn't buy that, my AI did'. Agentic commerce to drive new wave of disputes

'I didn't buy that, my AI did'. Agentic commerce to drive new wave of disputes

Chargebacks911 warns of new dispute patterns driven by the rise of agentic commerce

Chargebacks911, a company specializing in chargeback prevention and dispute management, has warned of new dispute patterns emerging from the rise of agentic commerce. CEO Monica Eaton says that as AI agents execute purchases on behalf of consumers, entirely new types of chargebacks are emerging.

Traditional chargebacks had clear causes such as unauthorized card use or non-delivery of goods. In agentic commerce, however, disputes arise when AI agents execute purchases exactly as instructed, yet consumers claim they didn't want those purchases. The card wasn't stolen, the merchant made no error, and the agent operated within its settings. Yet disputes still occur — a fundamentally new dynamic.

Agentic commerce is a form of commerce where AI agents autonomously execute everything from product search and comparison to purchase and payment. In 2026, this space has rapidly moved into practical deployment.

Visa has launched "Visa Intelligent Commerce," an API and partner program with over 100 partners. Positioning 2026 as the year of mainstream adoption, Visa is running pilots across Asia-Pacific and Europe. With DBS Bank in Singapore, Visa completed a real transaction at a restaurant using an AI agent.

Meanwhile, Mastercard is conducting pilots through its "Agent Pay" framework with Westpac in Australia and Axis Bank in India. In New Zealand, an AI agent completed an actual movie ticket purchase, with the issuer, merchant, and acquirer all able to identify the transaction as agent-initiated.

While this infrastructure development progresses, Datos Insights forecasts that chargeback volumes will increase 24% from 2025 to 2028, reaching 324 million cases globally. The expansion of agentic commerce is poised to accelerate this growth.

The Intent Verification Gap Created by "Click-Free" Transactions

Eaton identifies the core issue: the payments industry has long treated the "click" as proof of intent, but agentic commerce removes that click entirely. A new challenge emerges around how to prove intent in transactions where humans are not directly involved.

Multiple dispute scenarios can arise. AI agents automatically renewing subscriptions, booking travel based on calendar events that doesn't match customer preferences, reordering products that are no longer needed, or selecting alternative brands for cost optimization. Even when the system operates correctly, misalignment with consumer expectations creates disputes.

According to Chargeback Gurus' analysis, beyond AI agent misjudgments, multiple dispute categories are anticipated that are fundamentally different from traditional fraud — including cases where multiple AI agents place duplicate orders for the same product, or consumers fail to accurately remember the scope of authority they delegated to their agents.

Furthermore, agentic commerce is driving a shift toward performance-based pricing models. Fee structures tied to AI-completed purchases are emerging, and as background purchases increase, so does the risk of charges consumers don't recognize.

Traditional chargeback responses relied on evidence such as authentication checks and delivery confirmations. Agent-initiated transactions, however, require a new evidence framework: what the consumer authorized the agent to do, what restrictions were configured, what the agent actually executed, and when the consumer was notified.

Implications and Action Items for E-Commerce Businesses

Legal frameworks are also in urgent need of development. According to Torys LLP's analysis, existing legal frameworks cannot address the fundamental question of who bears liability when an AI agent executes a transaction — the user, the AI company, or the platform. The EU's AI Act will enforce strict rules in August 2026, with penalties of up to 7% of global revenue for violations. Colorado's AI Act also takes effect in June 2026, introducing consumer protection regulations.

The actions e-commerce businesses should take immediately are clear.

Clarifying agent permissions is the top priority. Businesses need to implement systems that allow granular control over the types of transactions AI agents can execute, including spending limits and category restrictions. They should also consider adopting standards like Visa's Trusted Agent Protocol that distinguish legitimate AI agents from malicious bots.

Enhancing transaction visibility is essential. Businesses must build real-time notifications, agent transaction flags in transaction details, and dashboard views so consumers can immediately identify agent-executed transactions. Mastercard's pilot has implemented a system where agent-initiated transactions are identifiable throughout the entire payment flow.

Building evidence trails is the cornerstone of dispute prevention. Businesses need centralized management of delegation scope, configured restrictions, execution logs, and notification history to present as evidence when disputes arise.

Conclusion

While agentic commerce has the potential to dramatically improve efficiency and convenience in the e-commerce industry, it is shaking the fundamental assumption that payments are based on "human intent." With Visa and Mastercard — the two major networks — accelerating their pilots, building new dispute resolution frameworks is an urgent priority for the entire industry.

The next key developments for e-commerce businesses to watch are the commercial rollout of agentic payments in the second half of 2026 and changes in the regulatory environment following the EU AI Act's enforcement. As Eaton emphasizes, agentic commerce only works when it places customer intent at the center of every transaction. Designing dispute prevention systems alongside technology adoption is an essential perspective for operating in this new era of e-commerce.

Related Articles

Tags

Agentic CommerceAINews

Start running your Shopify store smarter, today.

Connect Presso to Claude Code in under 10 minutes. Start your 14-day free trial with full access.

Start Free Trial