Most objections to voice AI are reasonable guesses about how people will behave. Having now analysed a large volume of real, consented travel calls, we can replace some of those guesses with evidence. Five of the most common myths, examined.
Myth 1: people hang up when they realise it's an AI
Some do, and they are entitled to. But the large majority continue when the assistant is helping them quickly. Hang-ups correlate far more strongly with slow responses and misunderstandings than with the caller learning they are speaking to an AI.
Myth 2: it only works for simple questions
Multi-step amendments — change the date, check the hotel still fits, add insurance, confirm the new price — are completed routinely when the assistant has access to the underlying systems. The limit is integration depth, not conversation complexity.
Myth 3: it will go off script and make things up
An assistant grounded in your knowledge base with explicit guardrails answers from your content or says it cannot help. The failures we see are almost always missing content, which is fixable, rather than invention.
Myth 4: older callers won't use it
Completion rates vary with clarity of speech and line quality far more than with age. Callers of every age complete tasks when the assistant speaks clearly, confirms details and does not rush.
Myth 5: it replaces the team
It changes the mix of work. Routine volume moves to the assistant; agents spend more time on complex, high-value and emotionally demanding calls. Headcount decisions are the business's to make — the technology simply removes the queue.
Design for speed, clarity and honesty and most of the myths take care of themselves.
