Will my customers know they're talking to AI?

Whether your customers know they're talking to AI depends on whether you tell them. Modern AI voice technology is natural enough that most callers can't determine it from the voice alone — but disclosure is increasingly required by law in certain U.S. states, and research consistently shows that proactive transparency builds rather than erodes trust.

What the law says

Several states and federal guidelines now address AI voice disclosure:

• California (B.O.T. Act, SB 1001) — requires disclosure when using a bot to influence a commercial transaction or election. Scope is narrower than many businesses assume, but applies to parts of sales and political communication.

• Illinois (HB 3773, AI video/audio disclosures) — requires disclosure for certain synthetic media interactions in employment contexts

• FCC — in 2024, ruled that AI-generated voices in robocalls require the same TCPA consent as other prerecorded voices, with new disclosure requirements pending as of the latest rulemaking

• Federal Trade Commission — has indicated that undisclosed AI in commercial conversations may be an unfair or deceptive practice under FTC Act Section 5

• International — the EU AI Act requires disclosure when users interact with an AI system; applicable if you serve European customers

The safe default: proactively disclose, except in contexts where disclosure itself would be misleading (automated IVR-style flows where the caller hasn't engaged in a conversation yet).

What disclosure actually looks like

Disclosure doesn't need to be clumsy. Options that work well:

• In the greeting — "Hi, this is the AI receptionist for Dr. Smith's office. How can I help?"

• On request — the AI openly confirms when asked "are you a real person?" or "am I talking to a computer?"

• At the point of sensitive action — "I'm an AI assistant; let me transfer you to someone on our team to discuss your concerns about this"

• In the follow-up text — confirmation messages can note "your call was handled by our AI receptionist"

What the research says about customer reactions

• Most callers don't mind interacting with AI as long as it's competent and the task is routine

• Callers react negatively when they feel deceived — for example, being confidently told "I'm just a regular receptionist" when they're not

• Disclosure combined with competence generally produces higher satisfaction than no disclosure

• Callers who prefer humans appreciate being told so they can ask to be transferred immediately

When customers may feel differently

• Healthcare and behavioral health — some patients feel strongly about speaking only to humans in clinical contexts; accommodate this

• Complaint or emergency calls — disclosure combined with immediate escalation to a human reduces friction

• High-value transactions — many consumers prefer human judgment for significant purchases or decisions; offer the human path

• Older callers — some demographics are more skeptical of synthetic voices; train your configuration to recognize hesitation and offer human transfer quickly

What we recommend

• Disclose proactively in most commercial voice AI deployments

• Make the disclosure short and natural, not a legal disclaimer

• Give the caller an obvious path to a human if they want one

• Monitor caller satisfaction for escalation patterns that suggest your disclosure or transfer options need tuning

Related concepts

• Do AI phone agents sound robotic?

• What is the TCPA?

• How does the AI escalate to a human?

• ISO/IEC 42001 AI management

See it in action

The Receptionist Agent at 365agents is designed for proactive disclosure by default. You can configure the exact wording to match your brand voice, and our compliance team helps you stay on top of evolving AI disclosure laws as they're passed.