Brands increasingly put AI in front of customers: shopping assistants, website chatbots, social reply agents and "brand concierges". Marketing often owns the experience — and the brand risk.
You are responsible for what your bot says
In February 2024, British Columbia's Civil Resolution Tribunal decided Moffatt v. Air Canada. The airline's website chatbot had told a customer he could apply for a bereavement fare after travelling; the actual policy said otherwise. Air Canada argued the chatbot was responsible for its own statements. The tribunal rejected that, holding that the chatbot was part of Air Canada's website and the company was responsible for all information on it. Air Canada was ordered to pay the customer damages.
What can go wrong
- Wrong facts: prices, policies, availability, eligibility.
- Off-brand tone: jokes, slang or arguments that go viral for the wrong reasons.
- Promises: discounts, refunds or commitments the business didn't make.
- Manipulation: users trying to make the bot say harmful or embarrassing things.
- Disclosure: customers may need to be told they are talking to an AI (for example under the EU AI Act's transparency rules for systems that interact with people).
Guardrails that work
- Ground answers in approved, current sources (policies, product data), and say "I don't know" otherwise.
- Limit scope: what topics it handles, and what goes to a human.
- Block promises the bot can't make (refunds, legal or medical advice).
- Red-team before launch: try to break it, including with hostile prompts.
- Monitor and log conversations; review samples weekly; keep a fast way to switch it off.
- Easy handoff to a human.