What legal, liability and fraud risks does agentic commerce create for brands?

By James Thomson, Managing Partner of Equity Value Advisors · Published October 12, 2026 · Last updated October 12, 2026

Short answer

Agentic commerce creates four families of risk for brands: misrepresentation, where AI restates product claims incorrectly; unauthorized or mistaken purchases, where liability among shopper, agent provider, merchant and brand is unclear; fraud, including agent impersonation and manipulated product data; and regulatory exposure as consumer protection, privacy and agent-access rules develop. Boards should confirm management has mapped each risk and has contracts, controls and dispute processes in place.

This page describes governance questions, not legal advice. Brands should involve counsel on specific obligations.

Key points

Companies remain responsible for what their own AI tools tell customers, as the 2024 Moffatt v. Air Canada decision showed.
No settled rule yet allocates liability when a third-party agent buys the wrong product or exceeds its authority.
Fraud will adapt to agent architecture. Treat agent transactions as a new transaction type, not as trusted automation.
Regulation is moving: the proposed AI AGENT Act would give the FTC a registration role over agent providers.

1. Misrepresentation: who is responsible when AI gets a product claim wrong?

When a brand’s own chatbot or agent misstates a product, price or policy, the brand should expect to be held responsible. In Moffatt v. Air Canada (February 2024), the British Columbia Civil Resolution Tribunal found Air Canada liable for negligent misrepresentation after its website chatbot gave a customer incorrect bereavement fare information. The tribunal rejected Air Canada’s argument that the chatbot was a separate entity responsible for its own statements, and held that correct information elsewhere on the site did not cure the error.

Liability for errors made by third-party assistants, such as ChatGPT or Gemini describing a brand’s product, is less settled. The reputational exposure is not: shoppers often blame the brand when a product fails to match what an AI told them. The defense is accurate, consistent source data and documented substantiation for every claim. See how to protect product accuracy in AI recommendations.

2. Unauthorized and mistaken purchases: who pays when an agent gets it wrong?

Agents may buy the wrong size, variant or quantity, purchase from an unintended seller, or exceed the spending limit the shopper set. The parties involved include the shopper, the agent provider, the payment provider, the merchant of record and the brand. Existing payment and consumer protection rules were written for human purchasers, and the AI AGENT Act, as introduced, does not clearly allocate liability among them.

Payment networks are building the evidence trail. Mastercard Agent Pay uses agentic tokens and requires agents to be registered and verified. Visa’s Trusted Agent Protocol helps merchants recognize trusted agents and receive information about agent intent and the consumer. These tools help establish who authorized what, but they do not by themselves settle who bears the loss.

Brands should decide in advance how they will handle returns, refunds and disputes on agent-placed orders, and make sure service teams can see when an order came through an agent. See who is responsible when an AI shopping agent buys the wrong product.

3. Fraud: how will bad actors exploit agentic commerce?

Fraud vector	How it works	Control
Agent impersonation	Malicious bots pose as legitimate shopping agents	Verify agent identity through network programs and bot management
Credential abuse	Stolen or over-broad payment credentials used by automated buyers	Tokenized, transaction-specific authorization and spending limits
Permission manipulation	Agents induced to exceed the shopper’s authority	Clear permission scopes and confirmation for high-value orders
Poisoned product data	False product information planted where agents read it, including prompt injection	Authoritative brand data, monitoring of AI answers and listings
Returns and refund abuse	Disputes that exploit unclear agent responsibility	Agent order flags, dispute evidence and consistent policies
Reseller manipulation	Unauthorized sellers alter listings to win agent routing	Authorized seller enforcement and listing monitoring

Automation favors convenience, and fraud controls add friction. The goal is controls that recognize legitimate agents without dropping verification, monitoring or customer recourse. See what fraud risks agentic commerce creates.

4. Regulatory exposure: what rules are emerging?

Agent access. The AI AGENT Act (S. 5051), introduced by Senator Mark Warner on July 21, 2026, would require covered platforms to give authorized user agents nondiscriminatory access, impose fiduciary-style duties on agent providers, bar them from using user data for advertising and give the FTC a registration and enforcement role. It is pending in the Senate Commerce Committee.
Platform access litigation. The Ninth Circuit’s August 2026 ruling in Amazon v. Perplexity held that an agent acting on a user’s instructions is the user for purposes of the Computer Fraud and Abuse Act. The underlying case continues.
Advertising and disclosure. Existing truth-in-advertising and endorsement rules still apply to product claims and paid placement, including placement inside AI shopping experiences.
Privacy. Sharing customer data with agents and platforms falls under existing privacy and consent laws. See what customer data brands should share with AI shopping agents.

What should boards ask about agentic commerce risk?

Have we mapped misrepresentation, purchase error, fraud and regulatory risks for each agent channel, with an owner for each?
Do our own AI tools draw only on substantiated, current product information?
Can we identify agent-placed orders, and do we have return and dispute policies for them?
Do our contracts with platforms, retailers and payment providers address liability for agent errors?
Which proposed rules would change our obligations, and who is tracking them?

Most of this work belongs to management. It becomes a board matter when it creates material financial, regulatory, concentration or reputational exposure. See the seven questions boards should ask about agentic commerce readiness.

What good looks like

Management maintains a short risk register for agentic commerce, reviewed with legal and finance. Product claims are substantiated and consistent across channels. Agent-placed orders are identifiable, with clear dispute handling. Contracts address agent errors, and someone tracks emerging regulation.

Frequently asked follow-up questions

Is a company liable for what its AI chatbot tells customers?

The leading decision says yes. In Moffatt v. Air Canada (2024), a Canadian tribunal held Air Canada liable for incorrect information its website chatbot gave a customer, rejecting the argument that the chatbot was a separate entity.

Who is liable when an AI shopping agent makes an unauthorized purchase?

As of October 2026, no settled rule allocates that liability among the shopper, agent provider, payment provider and merchant. Payment network programs aim to document authorization, and contracts and merchant policies fill the gap for now.

What is the AI AGENT Act?

The AI AGENT Act (S. 5051) is a US Senate bill introduced in July 2026 by Senator Mark Warner. It would require large platforms to give authorized AI agents nondiscriminatory access, impose duties on agent providers and give the FTC a registration and enforcement role. It is pending in committee.

Sources

American Bar Association, BC Tribunal Confirms Companies Remain Liable for Information Provided by AI Chatbot, February 2024.
Davis Wright Tremaine, The Federal AI AGENT Act, July 2026.
The Next Web, Court lets Perplexity’s AI agent shop on Amazon, overturning a ban, August 5, 2026.
Visa, Visa Introduces Trusted Agent Protocol, October 2025.
Mastercard, Mastercard unveils Agent Pay, April 2025.

What changed in agentic commerce, and what it means for brands.

A concise weekly brief for brand executives navigating the shift from human-led shopping to agent-mediated commerce. Follow the developments that matter, the implications for brand control and governance, and the questions senior teams should be asking now.

Read The Agentic Commerce Brief on Substack