Agentic Commerce and the Risk of Increasing Chargebacks

By Tim Tynan, CEO, Chargeback Gurus

In traditional e-commerce, a human makes the decision to buy. With agentic commerce, that decision may happen automatically based on customer preferences and signals.

For example, an AI agent might automatically buy household staples like detergent or pet food when supplies run low, based on sensor data or past purchase frequency. A customer might also ask the AI to purchase a type of product without specifying a particular brand, leaving the AI to decide which one to order.

Each scenario opens the door to a chargeback if the customer is dissatisfied with the purchase the AI agent made. For merchants, that may mean lost revenue, fees, and higher chargeback ratios that can trigger penalties or account restrictions.

Ideally, customers would contact the merchant directly for a refund if they have a problem with a purchase made by their AI agent. This allows the customer to get their money back more quickly and allows the merchant to avoid additional fees and penalties. Unfortunately, some consumers will always assume it’s easier to dispute the charge with their bank instead. This is especially true when the customer might not know which merchant the AI agent made the purchase from or how to contact them.

Some aspects of how chargebacks on AI-initiated transactions will be handled are still up in the air. Transaction records may not clearly show the customer’s consent, and banks are still defining what constitutes valid authorization when AI is involved. However, there are a few basic steps that can help merchants ensure they’re prepared:

1. Helpful billing descriptors: Ensuring the business name is clear and including a customer service phone number can help direct customers to contact the merchant instead of disputing the charge.

2. Confirmation emails: Sending a confirmation email with order details ensures the customer is aware a purchase has been made. If chargebacks due to unwanted purchases become more common, merchants may consider including a link to cancel the order in the email.

3. Additional verification: For higher-value or higher-risk orders, merchants can require additional verification to ensure the purchase is authorized by a human. Once available, tools like Visa’s Trusted Agent Protocol could also be used to verify trusted AI agents.

4. Investing in analytics: Identifying the root causes of chargebacks, whether connected to AI agents or something else entirely, allows merchants to make operational changes that reduce disputes.

Disputes linked to AI-driven transactions can be caused by already-common issues like customer dissatisfaction, or they could stem from something entirely new, such as AI misinterpreting a product name or description.

As agentic commerce grows, understanding the patterns behind chargebacks will be critical. Over time, analyzing chargeback data can reveal trends across products, customer segments, and even specific AI platforms. Specialized chargeback management companies, like Chargeback Gurus, can combine data from across a broad client base to detect emerging patterns in how issuers evaluate AI-related chargebacks and which forms of evidence are effective.

Chargeback Gurus is already working on methods of distinguishing between transactions made by humans and those made by various AI agents to help merchants identify whether these any of these agents are causing frequent chargebacks. Tools like Visa’s Trusted Agent Protocol and PayPal’s agentic payments system will no doubt be invaluable in this effort.

Data analysis can guide merchants in refining both their customer communication and their representment strategies, helping them adapt to the realities of AI-driven commerce. The merchants who thrive will be those that combine innovation with vigilance, treating every chargeback not just as a cost to recover, but as data to learn from.