Is a Voice Memo a Prerecorded Call? What Howard v. RNC Leaves Open

Why a voice memo is a legal question at all#

Thirty seconds of a rep's own voice does work a paragraph cannot, which is why voice memos show up in so many iMessage sales plays. The same feature is the one part of a text the prerecorded-voice rule can reach.

Section 227(b)(1)(A)(iii) of the TCPA bars making any call to a cell number using an artificial or prerecorded voice without the called party's prior express consent.

Courts treat a text as a call. Howard v. Republican National Committee reaffirmed that in January 2026 and said the answer would be the same without the old Chevron deference.

• Damages run $500 per violation under 47 U.S.C. 227(b)(3), and a court may triple the award for a willful or knowing violation (statute text, read 9 Oct 2026).
• Telemarketing raises the bar. 47 CFR 64.1200(a)(2) requires prior express written consent for a call that introduces an advertisement or is telemarketing and uses a prerecorded or artificial voice.
• Informational messages need prior express consent, which is lighter, but not zero.
• The FCC applied the same rule to ringless voicemail in FCC 22-85 (released 21 Nov 2022): a message dropped straight into a cell voicemail box is a call made using a prerecorded voice, so it needs consent.

What Howard v. RNC actually held#

On 24 October 2020 the RNC texted Jacob Howard's cell a few lines of copy, a link, and a video of Ivanka Trump. The video downloaded on its own but showed as a thumbnail with a play button.

Howard sued under section 227(b). The Ninth Circuit affirmed dismissal 2 to 1 on 13 January 2026 (No. 23-3826, opinion by Judge Collins, dissent by Judge Rawlinson). Read in full on 8 Oct 2026.

• The court agreed the text was a call and that the video contained a prerecorded voice. Neither point helped the RNC.
• The holding turns on two verbs. To make or initiate a call means how the call is begun, so the rule reaches prerecorded voices used to begin a call and nothing after that.
• Howard conceded he had to tap the thumbnail to hear anything. The court called that tap a necessary intervening action between the RNC's contact and the voice playing.
• The text, in the court's words, was made or initiated by its textual content and its silent inclusion of a ready-to-play video file.
• Footnote 8 dealt with autoplay. A phone set to play incoming video by itself did not change the result, and the court said a person who chooses that setting has given prior express consent to hearing it.

Why a standalone voice memo is a harder case#

Howard's text carried written words, with the video shown as a thumbnail above them. A voice memo sent on its own has no written message to begin the call with. The audio is the message.

The majority's own example points the other way. It said a prerecorded voicemail intrudes as soon as the person plays it back, and contrasted that with a recording a live speaker leaves at the time.

So the closer a memo looks to a voicemail, the less Howard's reasoning carries it. The closer it looks to a written text with optional media attached, the more it does.

• The dissent read the statute as covering any call that uses a prerecorded voice, citing Loyhayem (9th Cir. 2021) and the Supreme Court's line in Facebook v. Duguid: prerecorded voice calls irrespective of the type of technology used.
• Trim v. Reward Zone (9th Cir. 2023) had already said in footnote 4 that a text sent as MMS could include audio with an artificial or prerecorded voice. Howard confirmed the voice can be there; it ruled on when it is used.
• Howard binds federal courts in Alaska, Arizona, California, Hawaii, Idaho, Montana, Nevada, Oregon and Washington. A judge anywhere else can follow it, ignore it, or adopt the dissent.

Three kinds of voice memo, three levels of exposure#

Sales teams send voice memos in three different ways, and the law sees each one differently. Sort yours before deciding which consent you need.

The table reads the federal rules only. State statutes come after it.

AI voices are artificial voices#

The FCC settled the AI question in a Declaratory Ruling, FCC 24-17, adopted 2 February 2024 and released 8 February 2024. Read in full on 8 Oct 2026.

Voice cloning and similar tools fall inside the existing ban on artificial or prerecorded voice messages, the Commission said, because a person is not speaking them.

• Consent comes first. Callers must get prior express consent before a call that uses a voice simulated or generated by AI, and prior express written consent if it is telemarketing (footnote 13, citing 47 CFR 64.1200(a)(2) and (a)(3)).
• No live-agent carve-out. The ruling says the TCPA does not allow a carve-out for technologies that purport to provide the equivalent of a live agent.
• AI voice memos are on sale in this category. Project Blue's FAQ offers text-to-speech memos sent to hundreds of contacts without recording, plus cloning of the sender's own voice (tryprojectblue.com, 9 Oct 2026). Treat a cloned memo as an artificial-voice call, not as a rep's memo.

How an iMessage voice memo actually arrives#

Howard turned on what the recipient had to do, so the mechanics on the receiving phone matter. Apple's iPhone User Guide for iOS 27 describes them (read 8 Oct 2026).

Nothing in Apple's description plays audio on arrival. Every playback path starts with something the recipient does.

• Playback starts when the recipient taps play in the conversation, or raises the iPhone to their ear with Raise to Listen on. Raise to Listen can be turned off in Settings.
• Apple transcribes audio messages automatically, in the language chosen in the sender's settings.
• The sender's copy leaves the conversation 2 minutes after sending unless they tap Keep or set Expire to Never, and a recipient must tap Keep within 2 minutes of listening. Keep your own record of what was said.

Pennsylvania adds a state layer on 18 October#

Pennsylvania Act 47 of 2026 was signed on 20 July and takes effect in 90 days, on 18 October 2026. Its robocall definition is written for messages, not only calls.

The definition needs both pieces: an automated dialing system and a prerecorded or artificial voice. A memo a rep records and sends by hand fits it poorly; a workflow blast fits it better.

• A robocall is a telephone solicitation that uses an automated dialing system to deliver prerecorded or artificial voice calls or messages. Section 5(a)(11) then requires prior express written consent on any residential, business or wireless line.
• Business lines are covered by name, which the federal do-not-call claim does not do. The full Act 47 breakdown is at /blog/pennsylvania-act-47-text-messages.
• Section 5(a)(13) bars any synthetic or computer-generated messaging used to defraud, deceive or mislead a subscriber.
• Florida's FTSA, section 501.059(8)(a), needs prior express written consent for an unsolicited telephonic sales call that plays a recorded message when a connection is completed, but it covers consumer goods and services, so most B2B pitches sit outside it.

A voice memo policy a sales team can run#

None of this argues for dropping voice memos. A memo recorded for one person, after they wrote back, is a strong reply-rate move and the lowest-risk use of the feature.

Write the rules down once, so a workflow builder in your CRM cannot quietly turn a rep's memo into a broadcast.

• Open cold threads with text. The first message to a cold prospect should be written words; save the memo for after the reply. The same logic sits behind /blog/no-link-in-the-first-message.
• Record live, per prospect, for sales content. One file reused across a list is a prerecorded voice, and for sales content the safe plan is prior express written consent.
• Reuse a recorded memo only for people who asked for it, such as a pre-call agenda for a booked call or a reminder for webinar registrants.
• Start every reused memo by naming the business as registered with the state, and give a phone number. 47 CFR 64.1200(b) requires both for any artificial or prerecorded voice message.
• No cloned or synthetic voice without prior express written consent on file. FCC 24-17 leaves no gray area there.
• Keep the audio file and its transcript on the contact record. If a claim arrives, the question will be what the memo said and who heard it.

Where Blue Reacher sits in this#

Blue Reacher sends over iMessage, so no A2P registration is required. That is a carrier-registration fact, and it changes nothing about consent law or the prerecorded-voice rules.

We are a vendor in this category. Nothing here is legal advice, and every citation is dated so your counsel can check it.

• On the API, a voice memo is sent from an audio file you host and converted to Apple's native voice bubble when possible, so the file you sent is the file you keep. Details at /features/rich-media-messaging.
• Pacing is published: no total daily message cap, replies are never metered, and new conversations pace at 50 opted-in or 30 cold new contacts per line per day.
• For whether a text is a call in the first place, circuit by circuit, see /blog/are-texts-calls-tcpa-dnc-map.

Frequently asked questions