THEAGENCY.STRATEGY BUILDS WHAT’S NEXTStart a conversation →

Scale with AI / Episode 37 / November 9, 2026

Scale with AI — November 9, 2026

A familiar voice is no longer proof. AI can copy a voice from seconds of public audio and hold a live conversation in it, so the defense is not spotting the fake — it is checking the request: the call-back rule, two-person approval, and a code word. Plus the November 25 opt-out rules, sixteen days out.

The Agency LLC · 7 min · Updated November 9, 2026

Scale with AI podcast cover art

Listen. Then put it to work.

Read the full transcript ↓

Quick Take

  • The Shift: a familiar voice was proof for all of business history. That rule is finished — a voice is now a claim, and the same is true of a face on a video call and a flawless email with the right invoice attached.
  • The Law: sixteen days to the November 25 opt-out rules. A marketing opt-out ends all of your marketing calls and texts; opt-outs must be honored within ten business days at the latest; $500 to $1,500 per violation.
  • Who's Doing It: on November 5, federal prosecutors announced charges in an alleged fraud of about $129 million using AI copies of a chief executive's voice and face. The FBI logged more than 22,000 AI-related complaints in 2025, with losses above $893 million.
  • Your Move: write the call-back rule, pick your two-person limit, choose a code word and share it in person — then tell your team, out loud, that following the rules will never get them in trouble.

The Shift

For all of business history, a familiar voice was proof. If it sounded like the owner, it was the owner. That rule is now finished, and it is worth understanding why, mechanically, so you can protect your company properly. Modern voice technology can learn the sound of a person's voice from a very short sample of public audio. A voicemail greeting. A video on your company page. A podcast interview. From that sample, it can generate new sentences, in that voice, in real time, and hold a two-way conversation. It can answer questions. It can sound rushed, or apologetic, or secretive, on cue. The newest systems can even listen and speak at the same time, interrupting and reacting the way a real person does, which removes the awkward pauses that used to give the game away. So the takeaway is uncomfortable but simple: a voice is no longer evidence. It is a claim. And the same is now true of a face on a video call, and of a perfectly written email with the right invoice attached. Notice what this does to the training most businesses rely on. We teach our teams to spot the scam by looking for the flaw — the misspelled word, the strange email address, the robotic voice. This scam has no flaw to spot. The email reads perfectly, because a machine that writes perfectly wrote it. The voice sounds perfect, because it was built from the real person's own recordings. The invoice looks perfect, because it is a copy of a real one, with one quiet change: the bank details. If your defense depends on your bookkeeper noticing something feels off, you no longer have a defense. The technology is genuinely wonderful when it answers your phones and books your jobs, and we build with it every day. But in the wrong hands, it means the request itself can never again be trusted just because of who it sounds like. You have to check the request, not the voice.

The Fine Print

Sixteen days. On November 25, revised federal rules on opting out take effect, and they change how your business must handle a customer who says stop. Here is the plain version. If a customer opts out of one kind of informational message, say appointment reminders, you may limit that opt-out to that kind of message. But if a customer opts out of marketing, that ends all of your marketing calls and texts, full stop. You may name specific ways for customers to opt out, but only if you state them clearly everywhere and they actually work when a customer tries them. And every opt-out must be honored within ten business days at the very latest, though the smart businesses treat same day as the real standard. The exposure for getting this wrong is set in the law itself, at five hundred to fifteen hundred dollars per violation, and those numbers add up across a call list very quickly. One more rule owners forget. If a customer tells a live member of your team, in plain words, to stop calling, that counts, on its own, no form required. Your system has to hear it, record it, and honor it.

Who’s Doing It

On November fifth, federal prosecutors announced charges in a fraud case built entirely on this technology. Prosecutors say a small group targeted thousands of companies around the world over about a year and a half, using email, and then AI copies of a chief executive's voice, and in some cases the executive's face on video, plus a supposed lawyer, to persuade finance staff to wire money for a confidential deal. The reported total is about one hundred twenty-nine million dollars taken. The reported victims are not careless people. They include hospital systems, a bank, and an investment company. These are allegations, and the defendants have not been convicted, but the method is the lesson, and it scales down perfectly. The small business version is one urgent call, in the owner's voice, asking the bookkeeper to pay something quietly and quickly. National figures tell the same story. The federal bureau that tracks internet crime received more than twenty-two thousand AI related complaints in twenty twenty-five, with reported losses above eight hundred ninety-three million dollars. And the businesses getting it right? Every protected business runs the same unglamorous system, and it has nothing to do with spotting fakes. Nobody can reliably hear the difference anymore, and they have stopped trying. Rule one, the call-back rule: any request to move money, change bank details, or pay a new supplier gets confirmed by calling the person back on a number the business already had — never the number in the message, never a number the caller provides. A fake voice can call you. It cannot answer your real supplier's known phone. Rule two, two people: no single person can approve a new payee or a payment above a limit the owner sets, and urgency never skips the second person, because urgency is the weapon. Rule three, a code word: the owner and the team agree on one word for genuine urgent requests. No code word, no action, no argument. Three habits, written down, known by everyone who touches money.

Your Move

Three moves, all free, all doable before Friday. First, write your call-back rule in one sentence and tell your team: any change to payment details stops the payment until it is confirmed on a number we already trust. Second, pick your two-person limit — decide the amount above which two people must approve, and who the second person is. Third, choose your code word, share it in person, and never write it in an email. Then do one more thing, and this is the step that makes it real. Tell your team, out loud, that following these rules will never get them in trouble, even when the person asking sounds like you, and even when it turns out the request was genuine and the delay cost you an hour. The scam works because a good employee wants to be helpful and does not want to bother the boss. Your rules only protect you if your people believe you will thank them for checking. If a request arrives that is urgent, secret, and missing the word, your team now knows it is not a request. It is a test, and they just passed it. This is exactly the kind of protection we build into the systems we set up for businesses like yours — the answering, the follow-up, the reminders, with the consent rules, the opt-out handling, and the verification habits built in from the start, because the technology is only an asset when the rules around it are sound. If you want that working in your business, start with a <a href="https://theagencyllc.net/free-ai-audit/">free AI Audit</a> at The Agency. We will look at how your phones, your messages, and your payment requests actually flow, and show you where the gaps are.

Key Takeaways

  • A voice is no longer evidence — it is a claim. Modern systems can learn a person's voice from a very short sample of public audio and hold a two-way conversation in it, in real time, sounding rushed or secretive on cue.
  • This scam has no flaw to spot: the email reads perfectly, the voice sounds perfect, the invoice is a copy of a real one with the bank details changed. If your defense is a bookkeeper noticing something feels off, you no longer have a defense.
  • November 25 opt-out rules: a marketing opt-out ends all marketing calls and texts; every opt-out honored within ten business days at the latest; $500 to $1,500 per violation; a stop told to a live team member counts on its own.
  • Prosecutors announced charges November 5 in an alleged fraud of about $129 million built on AI copies of a CEO's voice and face; the FBI received more than 22,000 AI-related complaints in 2025, with reported losses above $893 million.
  • The protected businesses run three habits: the call-back rule (confirm on a number you already had), two people for any new payee or payment above a set limit, and one code word for genuine urgent requests — shared in person, never written in an email.

Full transcript

Alex: Welcome to Scale with A I for Monday, November ninth, twenty twenty-six. I am Alex.

Jordan: And I am Jordan. Today we are talking about the phone call you cannot trust anymore, even when the voice on the other end sounds exactly like your business partner, your spouse, or you. And more importantly, the simple system that makes that call harmless.

Alex: This is The Shift. For all of business history, a familiar voice was proof. If it sounded like the owner, it was the owner. That rule is now finished, and it is worth understanding why, mechanically, so you can protect your company properly. Modern voice technology can learn the sound of a person's voice from a very short sample of public audio. A voicemail greeting. A video on your company page. A podcast interview. From that sample, it can generate new sentences, in that voice, in real time, and hold a two-way conversation. It can answer questions. It can sound rushed, or apologetic, or secretive, on cue. The newest systems can even listen and speak at the same time, interrupting and reacting the way a real person does, which removes the awkward pauses that used to give the game away.

Jordan: So the listener's takeaway is uncomfortable but simple.

Alex: A voice is no longer evidence. It is a claim. And the same is now true of a face on a video call, and of a perfectly written email with the right invoice attached. Notice what this does to the training most businesses rely on. We teach our teams to spot the scam by looking for the flaw. The misspelled word, the strange email address, the robotic voice. This scam has no flaw to spot. The email reads perfectly, because a machine that writes perfectly wrote it. The voice sounds perfect, because it was built from the real person's own recordings. The invoice looks perfect, because it is a copy of a real one, with one quiet change, the bank details. If your defense depends on your bookkeeper noticing something feels off, you no longer have a defense. The technology is genuinely wonderful when it answers your phones and books your jobs, and we build with it every day. But in the wrong hands, it means the request itself can never again be trusted just because of who it sounds like. You have to check the request, not the voice.

Jordan: Which brings us to The Law, because November has a date in it that every business that calls or texts customers needs to know.

Alex: Sixteen days from now, on November twenty-fifth, revised federal rules on opting out take effect, and they change how your business must handle a customer who says stop. Here is the plain version. If a customer opts out of one kind of informational message, say appointment reminders, you may limit that opt-out to that kind of message. But if a customer opts out of marketing, that ends all of your marketing calls and texts, full stop. You may name specific ways for customers to opt out, but only if you state them clearly everywhere and they actually work when a customer tries them. And every opt-out must be honored within ten business days at the very latest, though the smart businesses we work with treat same day as the real standard. The exposure for getting this wrong is set in the law itself, at five hundred to fifteen hundred dollars per violation, and those numbers add up across a call list very quickly. One more rule owners forget. If a customer tells a live member of your team, in plain words, to stop calling, that counts, on its own, no form required. Your system has to hear it, record it, and honor it.

Jordan: Now Who's Doing It, and today that means who is being targeted, and who is getting it right.

Alex: On November fifth, federal prosecutors announced charges in a fraud case built entirely on this technology. Prosecutors say a small group targeted thousands of companies around the world over about a year and a half, using email, and then A I copies of a chief executive's voice, and in some cases the executive's face on video, plus a supposed lawyer, to persuade finance staff to wire money for a confidential deal. The reported total is about one hundred twenty-nine million dollars taken. The reported victims are not careless people. They include hospital systems, a bank, and an investment company. These are allegations, and the defendants have not been convicted, but the method is the lesson, and it scales down perfectly. The small business version is one urgent call, in the owner's voice, asking the bookkeeper to pay something quietly and quickly. National figures tell the same story. The federal bureau that tracks internet crime received more than twenty-two thousand A I related complaints in twenty twenty-five, with reported losses above eight hundred ninety-three million dollars.

Jordan: And the businesses getting it right? What do they actually do differently?

Alex: Every protected business we see runs the same unglamorous system, and it has nothing to do with spotting fakes. Nobody can reliably hear the difference anymore, and they have stopped trying. Rule one, the call-back rule. Any request to move money, change bank details, or pay a new supplier gets confirmed by calling the person back on a number the business already had, never the number in the message, never a number the caller provides. A fake voice can call you. It cannot answer your real supplier's known phone. Rule two, two people. No single person can approve a new payee or a payment above a limit the owner sets. Urgency never skips the second person, because urgency is the weapon. Rule three, a code word. The owner and the team agree on one word for genuine urgent requests. No code word, no action, no argument. That is the whole system. Three habits, written down, known by everyone who touches money.

Jordan: That is Your Move, so let us make it concrete for this week.

Alex: Three moves, all free, all doable before Friday. First, write your call-back rule in one sentence and tell your team. Any change to payment details stops the payment until it is confirmed on a number we already trust. Second, pick your two-person limit. Decide the amount above which two people must approve, and who the second person is. Third, choose your code word, share it in person, and never write it in an email. Then do one more thing, and this is the step that makes it real. Tell your team, out loud, that following these rules will never get them in trouble, even when the person asking sounds like you, and even when it turns out the request was genuine and the delay cost you an hour. The scam works because a good employee wants to be helpful and does not want to bother the boss. Your rules only protect you if your people believe you will thank them for checking. If a request arrives that is urgent, secret, and missing the word, your team now knows it is not a request. It is a test, and they just passed it. This is exactly the kind of protection we build into the systems we set up for businesses like yours, the answering, the follow-up, the reminders, with the consent rules, the opt-out handling, and the verification habits built in from the start, because the technology is only an asset when the rules around it are sound. If you want that working in your business, contact The Agency for a free A I audit. We will look at how your phones, your messages, and your payment requests actually flow, and show you where the gaps are. I am Alex.

Jordan: And I am Jordan. Check the request, not the voice. We will see you tomorrow on Scale with A I.

Sources

  1. U.S. Attorney's Office, Southern District of New York — charges announced November 5, 2026 ↗
  2. FBI Internet Crime Complaint Center (IC3) — AI-related complaints and reported losses ↗
  3. FCC — unwanted calls and texts, consent and opt-out rules ↗
  4. The Agency — Free AI Audit ↗

Your next conversation

Get Your Free AI Audit

Start a conversation ↗