Are AI Voicemail Drops Legal? Ringless Voicemail and the TCPA
Ringless voicemail is a robocall under the FCC 2022 ruling, and AI voices are covered too. What consent you need, the $500 to $1,500 per message exposure, and safer options.
By the ColdCalls.ai team
July 2026 · 9 min read
Ringless voicemail drops are not a TCPA loophole. On November 14, 2022, the FCC issued a Declaratory Ruling (FCC 22-85) holding that ringless voicemail messages delivered to a wireless number are "calls" under the Telephone Consumer Protection Act, so they require the same prior express consent as any other prerecorded or artificial-voice robocall. An AI-generated voicemail is treated the same way. Dropping marketing voicemails onto cold wireless numbers without consent exposes you to statutory damages of $500 to $1,500 per message.
This is a plain-English explainer for United States sales teams, not legal advice. TCPA is a private-right-of-action statute with an active plaintiffs' bar, and state law adds a second layer. Run your program past counsel before you scale it.
What the FCC actually decided
Vendors selling ringless voicemail spent years arguing that because the technology deposits a message directly into a voicemail server without causing the handset to ring, it never places a "call" to the wireless number and therefore sits outside the TCPA. The FCC rejected that argument outright. The 2022 ruling concluded that ringless voicemail delivery to a wireless number is a call made using an artificial or prerecorded voice, and that the sender needs prior express consent unless an exemption applies.
Two practical consequences follow. Consent has to exist before the message goes out, not after. And for messages that are marketing in nature, the standard is the higher one: prior express written consent, with the usual disclosure and signature requirements, not just a phone number someone typed into a form for a different purpose.
| Scenario | Consent needed | Practical read |
|---|---|---|
| Marketing voicemail drop to a cold wireless number | Prior express written consent | Do not do this on a purchased or scraped list |
| Voicemail drop to an existing customer about their service | Prior express consent, non-marketing treatment may apply | Depends heavily on the content of the message |
| Voicemail left by a live human rep after a manual dial | Outside the prerecorded-voice rules | Still subject to DNC and calling-hour rules |
| AI agent leaves an artificial-voice message on a cold wireless number | Prior express written consent for marketing | Same rules as any prerecorded drop |
Are AI voicemail drops treated differently from recorded ones?
No, and if anything the ground has moved further against them. The TCPA's language covers an "artificial or prerecorded voice," which plainly captures a synthesized AI voice. In February 2024 the FCC issued a separate Declaratory Ruling confirming that AI-generated voices in calls to consumers fall within that same artificial-voice prohibition. So a voicemail your AI agent composes and speaks on the fly gets no better treatment than a WAV file recorded by a rep in 2019.
State law tightens it further. Several states now require affirmative disclosure that a caller is an AI, and some regulate synthetic voices in commercial communications directly. A message that never identifies itself as AI is a compliance problem in those jurisdictions independent of the federal consent question.
What happens if you drop voicemails without consent?
TCPA damages are statutory and per message: $500 each, trebled to $1,500 for willful or knowing violations, with no cap. A campaign of fifty thousand drops to non-consented wireless numbers is theoretically a seventy-five million dollar exposure. That arithmetic is exactly why TCPA class actions are a functioning industry, and why the plaintiffs' bar watches voicemail-drop vendors closely. If you want to see how courts have actually applied the rules to specific fact patterns, the case law on the question is more instructive than any vendor's compliance page.
So when can you legally leave a voicemail?
Three situations are workable for most sales teams.
A live agent reaches voicemail on a permitted call. If the call itself is compliant, meaning the number is scrubbed against the federal and applicable state Do Not Call registries, you are inside permitted hours, and you have whatever consent the call requires, then leaving a message is generally an extension of that call rather than a separate campaign. This is the normal path for B2B outbound to business lines.
You have genuine written consent. A prospect who checked a properly worded box agreeing to receive automated calls and texts at the number they gave you is a different legal animal from a name on a purchased list. Keep the record of consent, the exact language shown, and the timestamp.
An established business relationship on a non-marketing message. Service, appointment and transactional messages to existing customers get more room than marketing does, but the content controls the classification. Slipping an offer into an appointment reminder converts it into a marketing message.
Does revoking consent apply to voicemail drops?
Yes. FCC rules that took effect in April 2025 require callers to honor a consumer's revocation of consent through any reasonable method, and to stop within ten business days. That means a "stop" spoken to your AI agent, texted back, or left on your line has to propagate to the voicemail-drop list, not just the dial list. If those two lists live in different systems and sync weekly, you have a defect that produces exactly the fact pattern plaintiffs look for.
The B2B nuance most teams get wrong
People often assume business-to-business calling is exempt from TCPA. It is not, and the assumption gets expensive when the "business number" turns out to be a cell phone. Sole proprietors, contractors, agents, franchise owners and most small-business decision makers use a wireless number as their business line. The TCPA follows the number, not the job title. If it is a wireless number, the prerecorded-voice consent rules apply regardless of who answers.
This is the single most common reason a B2B outbound program that felt safe on paper generates a demand letter. Before any automated voicemail campaign, run the list through a wireless identification check and treat every wireless hit as consumer-grade for compliance purposes.
How we handle it
Our position is straightforward: we do not run mass ringless voicemail campaigns to non-consented numbers, because the ruling is clear and the exposure is not worth it. The AI dialer places real calls, discloses on every connect that it is an AI, scrubs against DNC in real time, respects calling hours in the prospect's own time zone, and logs every attempt and disposition. When a call reaches voicemail on a permitted dial, the agent can leave a short disclosed message as part of that call rather than as a separate blast, and a revocation suppresses the number immediately across the whole program.
The broader framework, consent, disclosure, DNC, hours and record keeping, is laid out in our guide to whether AI cold calling is legal. If you are setting a program up from scratch, the compliant AI cold calling playbook walks through list hygiene and the disclosure script in order.
The short version
Ringless voicemail is a robocall under federal law, AI voices are covered by the same artificial-voice rules, and the exposure is $500 to $1,500 per message. Do not build a growth plan on voicemail drops to cold wireless numbers. Build it on compliant dialing, real disclosure and clean lists, then let the voicemail be a byproduct of a call you were allowed to make anyway.
See ColdCalls.ai book meetings
The AI SDR calls every lead, discloses it is an AI, qualifies, handles objections and books meetings into your calendar and CRM. Flat fee, no per-meeting cut, compliance built in.