A Canadian tribunal ruling set a precedent every sales organization should understand—if AI speaks on your behalf and gets it wrong, the liability is yours
.png)
Estimated Read Time: 4 minutes
In November 2022, a customer named Jake Moffatt lost a family member and booked an urgent flight. Air Canada’s chatbot told him he could apply for a bereavement fare retroactively within 90 days. He paid full price, submitted the request, and was denied. The airline’s position: the chatbot had been wrong, the correct policy was posted elsewhere on the site, and the chatbot was essentially its own system—not the airline’s responsibility.
The tribunal rejected this entirely. Its ruling was blunt: “It makes no difference whether the information comes from a static page or a chatbot.” Air Canada was ordered to pay $812 CAD in damages and fees. The dollar figure was small. The principle was not. Any organization deploying AI in client-facing communication is responsible for what that AI says—accurate or not.
The Air Canada case involved a customer service chatbot. But the principle extends to every AI-generated touchpoint in a sales organization. An AI tool that summarizes a discovery call and gets the prospect’s budget figure wrong. An AI email assistant that drafts a follow-up referencing a commitment the rep never made. An AI-generated proposal that lists a product capability your team no longer offers. Each of those outputs represents your organization to your prospect—and if the prospect acts on it, “the AI got it wrong” is not a defense.
McKinsey’s 2024 Global Survey on AI found that 44% of organizations reported at least one negative consequence from generative AI deployments—and by 2025, that number had climbed to 51%. These are not edge cases. They are the documented, recurring cost of deploying AI output without a verification layer. The question that should concern every sales leader is not whether their AI tools produce errors—it is whether those errors are being caught before they damage client relationships. The answer in most organizations is: not consistently. Explore the case for better data capture systems and what consistent, human-verified input changes across the full revenue stack.
The Air Canada failure happened because a system was deployed without meaningful quality control on its outputs. The chatbot had no mechanism for flagging uncertainty, no human reviewer catching edge cases, and no process for reconciling what it said against what the actual policy was. That absence was the liability—not the technology itself.
Sales organizations face an identical risk whenever AI-generated outputs reach clients or populate CRM records without review. An AI note that misquotes a pricing conversation. An AI summary that invents a follow-up commitment. An AI-generated competitive analysis that hallucinated a competitor’s feature set. Every one of these is an Air Canada waiting to happen—a confident, plausible, wrong statement that your organization owns the moment a client reads it. The voice-to-CRM systems that high-performing sales teams use are built around this reality: AI handles speed and structure; trained humans handle accuracy and accountability.
