Board Case Study · Consumer (Airline & Travel)
A support chatbot invented a refund policy, and a tribunal ruled the company must honor it. The same model has made thousands of unaudited promises. What else did it commit the company to?
The situation
A public travel company deployed a generative-AI customer-service agent to handle millions of interactions. A customer was told by the bot about a bereavement-refund policy that does not exist; when the company refused, a tribunal held that the company is bound by what its AI agent stated.
The ruling reframes every prior interaction as a potential binding commitment. The model has handled an enormous volume of chats, and no one has audited what it promised, fare rules, refunds, compensation, waivers. Each hallucinated assurance is now potential contractual or consumer-protection exposure.
The board must decide on retroactive exposure, guardrails on what the agent may state, human review for high-stakes commitments, and disclosure, while the tool remains central to a cost-reduced service model the company has already banked.
“A court just told us we're bound by what our chatbot says. It has answered millions of times, what did it promise while no one was reading?”
General Counsel
The decision on the table
The full case, scenario architecture, board materials, and facilitation notes, is shared with boards and partners on request.