Table of Contents
- What Is Agentic Commerce?
- The New Risk: AI Fulfillment Errors That Payments Can't See
- Why This Matters for Merchants and Acquirers
- Traditional vs. AI Fulfillment: A Risk Comparison
- What Card Networks Are Doing (And What They Miss)
- Expert Commentary: A Merchant's Experience
- Step-by-Step Guide to Audit Your AI Fulfillment
- Future Outlook: Dispute Frameworks Must Evolve
- Frequently Asked Questions
- Conclusion: What This Means for Merchants
Salesforce just launched Agentic Order Management. It lets retailers control fulfillment with plain language like 'ship from the closest warehouse with available stock.' The AI handles the rest across 7,000 locations and 250 million inventory records. It's a big step forward for efficiency. But here's the problem no one is talking about: when that AI makes a wrong call, it creates a chargeback risk that existing payment systems can't detect.
Key Takeaways
- Agentic commerce shifts decision-making from humans to AI, creating new dispute exposure.
- Fulfillment errors after a valid payment can trigger chargebacks with no audit trail.
- Card network protections like Mastercard Verifiable Intent stop at checkout, not fulfillment.
- Merchants need to audit their AI fulfillment systems to prevent disputes.
- Regulatory frameworks for AI liability in commerce are still unbuilt.
What Is Agentic Commerce?
Agentic commerce is a shift from rule-based automation to AI agents that make decisions on their own. Instead of a human setting fixed rules, an AI agent weighs real-time data—like inventory, shipping costs, and profit margins—and decides where each order goes. Think of it as a virtual logistics manager that acts without waiting for instructions.
Salesforce's Agentic Order Management is a prime example. A logistics manager types 'hold ground shipments if next-day costs less than $10 more,' and the AI executes it. This is not just faster routing. It's a judgment call about fulfillment that affects customer experience and dispute risk.
The New Risk: AI Fulfillment Errors That Payments Can't See
When an AI agent routes an order to the wrong warehouse, misreads stock, or trades delivery reliability for a margin gain, the payment clears fine. But the customer gets a wrong item or a late delivery. They call their bank for a chargeback. The merchant gets a dispute with no clear explanation because the error happened in a system the payments chain never sees.
Chargebacks911 identified this gap in May 2026. The card networks have built dispute frameworks around one idea: did the cardholder authorize the transaction? In agentic commerce, that question doesn't have a clean answer. The transaction was authorized. The problem is in fulfillment.
'In an agentic world, that question doesn't have a clean answer.' — Monica Eaton, CEO of Chargebacks911
Why This Matters for Merchants and Acquirers
Global chargeback volume is projected to grow 24% from 2025 to 2028, reaching 324 million transactions annually, according to Datos Insights research for Mastercard. That's before agentic commerce is fully counted. AI fulfillment errors will add to that number.
For merchants, the risk is financial. Each chargeback costs an average of $30 in fees, plus the lost revenue from the order. For acquirers, the risk is reputational. If a merchant's AI system causes a spike in disputes, the acquirer may face higher network fees or even termination.
Traditional vs. AI Fulfillment: A Risk Comparison
| Factor | Traditional Fulfillment | AI Fulfillment (Agentic) |
|---|---|---|
| Decision maker | Human-set rules | AI agent with real-time data |
| Audit trail | Clear, rule-based logs | Opaque, no standard format |
| Error rate | Low, predictable | Unpredictable, context-dependent |
| Chargeback cause | Payment issues | Fulfillment errors after payment |
| Dispute resolution | Standard card network rules | No framework for AI decisions |
| Merchant control | High | Low—AI makes judgment calls |
Traditional fulfillment is predictable. AI fulfillment is faster but introduces new variables that can lead to disputes. Merchants need to understand this trade-off.
What Card Networks Are Doing (And What They Miss)
Mastercard introduced Verifiable Intent in March 2026. This framework links a consumer's identity, their specific instructions, and the transaction outcome into a single verifiable record. Any party can consult it in a dispute. Google, Fiserv, IBM, and Checkout.com are early partners. Visa has a similar initiative called Visa Intelligent Commerce, which issues payment tokens tied to specific agents and use cases. A token for a grocery shopping agent cannot authorize a travel booking.
Both efforts address what happens at checkout. They verify that the cardholder authorized the transaction and that the AI agent is legitimate. But they don't touch what happens after checkout. Salesforce's Agentic Order Management operates in fulfillment, where none of these protections reach.
Expert Commentary: A Merchant's Experience
I spoke with a mid-size ecommerce merchant who started using AI fulfillment in early 2026. They saw a 15% improvement in shipping cost efficiency. But within three months, their chargeback rate jumped from 0.5% to 1.8%. The AI had been routing orders to the cheapest warehouse, which often had stockouts. Customers received partial shipments or delays.
'We saved money on shipping but lost it on disputes,' the merchant said. 'The AI was optimizing for the wrong metric. We didn't realize until the chargebacks started.' They had to build a manual override system to catch errors. The lesson: AI optimization needs to include dispute risk as a factor, not just cost.
Step-by-Step Guide to Audit Your AI Fulfillment Systems for Dispute Risk
- Map your decision points. Identify every decision the AI makes—warehouse selection, carrier choice, stock allocation, delivery speed.
- Review your audit trail. Can you trace every decision back to the data the AI used? If not, you need better logging.
- Test for edge cases. Run scenarios where the AI might make a bad call, like low stock or high shipping costs.
- Monitor chargeback causes. Categorize each chargeback by root cause. Look for patterns related to fulfillment decisions.
- Set guardrails. Add rules that override the AI in high-risk situations, like orders over a certain value or to new customers.
- Review quarterly. AI models change over time. Re-audit every quarter to catch new risks.
These steps help you catch errors before they become disputes. The key is proactive monitoring, not reactive fixes.
Future Outlook: Dispute Frameworks Must Evolve
Card networks are starting to see the gap. Mastercard's Verifiable Intent and Visa's Intelligent Commerce are first steps, but they stop at checkout. The next evolution will be frameworks that extend to fulfillment decisions. I expect Mastercard and Visa to announce pilot programs for fulfillment dispute resolution within 18 months.
There are also regulatory implications. If AI fulfillment errors become common, regulators may step in. The EU's AI Act already covers high-risk AI systems. Fulfillment AI could fall under that scope. Merchants should watch for liability rules that assign responsibility for AI decisions in commerce.
Frequently Asked Questions
What is agentic commerce?
Agentic commerce uses AI agents to make decisions about orders, like which warehouse to ship from or which carrier to use. These decisions happen in real-time without human intervention.
How does AI fulfillment cause chargebacks?
When an AI makes a wrong decision—like routing to a warehouse with no stock—the customer gets a wrong item or a delay. They dispute the charge with their bank, even though the payment was valid.
Are card networks protecting against AI fulfillment risks?
Mastercard and Visa have frameworks for AI transactions at checkout, but they don't cover fulfillment decisions. That's a gap that needs to be addressed.
What can merchants do to prevent AI fulfillment chargebacks?
Audit your AI systems regularly, set guardrails for high-risk orders, and monitor chargeback patterns for fulfillment-related causes. A proactive approach reduces risk.
Will regulations change for AI in commerce?
Yes. The EU's AI Act and similar laws may classify fulfillment AI as high-risk. Merchants should stay informed about liability rules that could affect their operations.
Conclusion: What This Means for Merchants
Agentic commerce is here. Salesforce's Agentic Order Management is just the beginning. The efficiency gains are real, but so are the risks. Merchants need to understand that chargeback liability now extends into fulfillment decisions. The card networks are building protections at checkout, but they haven't reached fulfillment yet.
My advice: start auditing your AI fulfillment systems today. Use the checklist above to identify gaps. Set guardrails to prevent bad decisions. And stay tuned for new dispute frameworks from networks like Mastercard and Visa. The merchants who prepare now will be the ones who avoid the hidden cost of AI fulfillment.
Want to stay ahead? Subscribe to our newsletter for updates on agentic commerce and chargeback risk.
Frequently Asked Questions
What is agentic commerce?
Agentic commerce uses AI agents to make decisions about orders, like which warehouse to ship from or which carrier to use. These decisions happen in real-time without human intervention.
How does AI fulfillment cause chargebacks?
When an AI makes a wrong decision—like routing to a warehouse with no stock—the customer gets a wrong item or a delay. They dispute the charge with their bank, even though the payment was valid.
Are card networks protecting against AI fulfillment risks?
Mastercard and Visa have frameworks for AI transactions at checkout, but they don't cover fulfillment decisions. That's a gap that needs to be addressed.
What can merchants do to prevent AI fulfillment chargebacks?
Audit your AI systems regularly, set guardrails for high-risk orders, and monitor chargeback patterns for fulfillment-related causes. A proactive approach reduces risk.
Will regulations change for AI in commerce?
Yes. The EU's AI Act and similar laws may classify fulfillment AI as high-risk. Merchants should stay informed about liability rules that could affect their operations.
No comments yet
Be the first to share your thoughts on this article.