Citi Says AI Companies Should Pay for Booking Errors
Citi argues that AI companies should bear financial liability for booking mistakes made by their agents, citing a Visa study showing a 43% error rate.
Stock photo for illustration only, not from the actual event
- Citi believes AI creators should absorb the liability for agent booking errors.
- A Visa study found 43% of tested agents made mistakes, with half completing the purchase anyway.
- Agent-led search drives 40% to 50% higher conversion and increases transaction spend.
- Citi is currently building its own in-app AI travel concierge.
As autonomous artificial intelligence agents increasingly handle travel bookings, the travel and financial industries are grappling with an unresolved question: who absorbs the financial cost when these systems make a mistake? Citi has taken a firm stance, arguing that the companies developing these AI agents should be held accountable.
Speaking at the Skift Commerce Summit in New York on Wednesday, Abhinav Anand, Citi's head of value cards, lending and commerce for U.S. consumer cards, emphasized that liability should rest squarely on the technology creators rather than shifting to hotels, airlines, or card issuers.
Stock photo for illustration only, not from the actual event
According to Citi, one viable enforcement mechanism is requiring AI agents to accept a degree of liability as a prerequisite for transacting on payment networks like Visa or Mastercard. This registration requirement could help protect merchants and consumers from runaway errors in automated travel planning.
Anand pointed to a Visa study involving more than 75 agents, which revealed that 43% made mistakes, and roughly half of those errors proceeded to complete the purchase. Even top-performing systems, which Anand noted included Google’s Gemini and Meta’s Muse, achieved an accuracy rate of about 82%, pointing to higher-than-average short-term chargeback rates.
"Liability should fall to the agent themselves, the company that is actually coming with the agent."
Abhinav Anand, Citi
The rise of agentic commerce introduces complex legal and operational challenges for traditional financial institutions and travel merchants. To resolve disputes fairly, the industry is currently working on embedding original traveler purchase instructions inside tokenized transactions, allowing issuers and merchants to verify intent during chargeback claims.
Despite the risks of errors, early industry metrics demonstrate immense commercial upside. Data indicates that agent-driven discovery increases conversion rates by 40% to 50% while lifting average spend per transaction. Capitalizing on this shift, Citi—an early adopter of agentic payments with Mastercard and Visa—is developing its own proprietary AI travel concierge inside its mobile application.
Source: Skift
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment