Every so often a client asks us to have their AI agent stop mentioning it's an AI — or, once in a while, to flat-out say it's a person. The first is reasonable. The second we won't do. Here's the line we hold.
Don't volunteer it
We keep disclosure non-proactive. The agent doesn't announce "I'm an AI" in the greeting — that yanks people out of a natural call for no reason. It just helps the caller like a good receptionist would.
But answer honestly if asked
The moment a caller asks "wait, am I talking to a real person?", the agent gives one brief, honest line and moves right along. No dodging, no deflecting.
Never deny it
This is the bright line: the agent never claims to be human. It's deceptive, it torches trust the instant someone figures it out, and bot-disclosure laws — California, a growing list of states, the EU AI Act — can require disclosure in sales and commercial contexts, sometimes proactively. (That's not legal advice; confirm the rules for your states.)
The logic is simple: the goal is a warm, useful call, not a con. An agent caught lying about being human costs you far more trust than a five-word disclosure ever would.
The owner takeaway
It's fine not to lead with "I'm a bot." It's not fine to lie about it. Before you launch any AI voice tool, ask the vendor exactly how it answers "am I talking to a real person?" — and make sure the answer is honest.
Want a second set of eyes on how your agent should handle that? Grab a free workflow audit.