Would Your Staff Catch a Cloned Voice or Video Call? Test It for Free
Most business owners think they’d know if the voice on the phone wasn’t really their CFO. A recent survey of cybersecurity leaders suggests that confidence is misplaced — and that the only way to find out for sure is to test it yourself, for free, before a real attacker does it for you.
What happened
A 2025 Gartner survey of 302 cybersecurity leaders, cited by security vendor IRONSCALES in its report “Deepfakes and the Confidence Problem,” found that 62% of those leaders said their organization had experienced at least one deepfake attack in the prior 12 months. Of those incidents, 43% involved a cloned voice on a phone call and 37% involved a faked video call.
A separate January 2026 Osterman Research study of 128 US security and IT leaders — also commissioned by IRONSCALES — found that 60% of those leaders doubted their own staff could reliably catch a deepfake attack, even after the training their companies already run.
Put those two numbers together and the picture is blunt: most security leaders have already had it happen, and most of them don’t trust their own defenses against it happening again.
Why this matters to a business like yours
“I recognized the voice” used to be a reasonable way to confirm who you’re talking to. It no longer is. AI voice-cloning tools can now reproduce a specific person’s voice, cadence, and word choice from a short sample of audio — a few minutes pulled from a company all-hands video, a podcast interview, or even voicemail greetings are often enough. The same is increasingly true of video.
You don’t need to be a large enterprise to be a target. The attack doesn’t need to fool a machine — it needs to fool one employee, for one call, long enough to get a wire transfer approved, a password reset pushed through, or an access badge reissued. Smaller companies are often easier targets precisely because they have fewer formal verification steps in place.
The uncomfortable part of the Osterman number isn’t that security leaders lack confidence — it’s that most of their companies have never actually tested whether that lack of confidence is justified. They’re guessing. You don’t have to.
What to do about it: run the drill yourself
A recent how-to published on The Hacker News lays out a version of this test that costs nothing but a few hours of staff time. You don’t need special software, a security vendor, or a budget line to run a first pass. Here’s the shortened version for a small or mid-sized business:
- Make two lists. Who’s worth impersonating — your CEO, CFO, or anyone who can approve money, resets, or access — and who’s likely to get the call: accounts payable, your IT help desk, whoever handles password resets.
- Write your verification rule on one page. A callback always goes to a number already in your internal directory, never a number the caller gives you. Set a dollar or action threshold above which a second person must sign off. And say out loud, in writing, that any employee is allowed to pause a request and go verify it — without being treated as difficult for doing so.
- Run one realistic fake scenario. A spoofed email followed by a cloned-voice call asking for a wire transfer, a password reset, or an access change — the way a real attempt would actually be chained together. Use a test account. Tell the people involved afterward it was a drill, not a trap.
- Debrief without blame, and repeat it twice a year. The point isn’t to catch someone failing — it’s to see where the process breaks: how long verification took, whether anyone paused to check, where the one-page rule wasn’t followed. Fix that, then run it again in six months.
That’s it. No purchase required — the method works with a notepad, a phone, and a willing coworker to play the attacker. If you later decide you want automated detection layered on top of your own process, that’s a separate decision for a separate day; the drill itself is something any business can run this month.
The bottom line
Deepfake voice and video attacks aren’t a future problem — the Gartner-cited survey says 62% of security leaders have already dealt with one. The honest question isn’t whether your company could be targeted. It’s whether anyone on your staff has ever been tested against a realistic attempt, under low stakes, before a real one shows up. A two-list, one-page, one-afternoon drill is the cheapest way to find out.
For more on how AI is changing both the risks and the tools available to ordinary businesses, see AIWAS’s knowledge base, and follow @AIWASai on X for the next alert.
Stop leaving customers on read.
We answer your reviews, keep your social pages posting, and make your ads convert, so you get more customers while you run the business.
Get Started