TCPA Compliance for AI Voice Agents: A Practical Guide

The Telephone Consumer Protection Act (TCPA) is the primary US regulation governing automated and prerecorded calls. Violations carry penalties of $500–$1,500 per call — and with AI agents capable of making thousands of calls per day, non-compliance risk scales rapidly. Understanding TCPA requirements isn't optional; it's existential for any outbound AI calling program.

Core TCPA requirements

• Prior express consent — required for automated calls to mobile phones. Written consent required for telemarketing calls.
• Do Not Call compliance — maintain and honor an internal DNC list; scrub against the National DNC Registry
• Calling hours — no calls before 8am or after 9pm in the callee's local time zone
• Caller identification — must transmit accurate caller ID information
• Opt-out mechanism — must provide an immediate way to opt out during every call
• AI disclosure — FCC rules require disclosure that the caller is AI-generated at the beginning of the call

Platform-level enforcement

Compliance cannot rely on prompt instructions. The LLM might skip a disclosure or fail to honor an opt-out under certain conversation conditions. These rules must be enforced at the platform level: hard-coded time-of-day restrictions, automated DNC list scrubbing before dial, mandatory AI disclosure audio at call start, and system-level opt-out processing that removes numbers regardless of what the agent says.

State-level variations

Several states have stricter rules than federal TCPA. California's CCPA adds data deletion rights. Florida and Washington have expanded private right of action for automated calls. New York requires specific disclosures for AI-generated voices. A national AI calling campaign needs compliance mapping across all target states — not just federal rules.