Savi のアプリは、身代金を要求するリアルな AI スキャンから消費者を保護することを目指している

Brothers Patrick and Ryan Coughlin, each with impressive careers in the tech industry (Patrick worked in national cyber defense, and at Splunk and Cisco; and Ryan with consumer products at Apple and Spotify), have launched a new kind of security startup.

Savi Security seeks to protect everyday folks from the new crop of incredibly convincing AI-generated scams, whether they’re routed via text, emails, or phone calls.

The company just raised $7 million in seed funding, and is launching its app for iPhone and Android on Tuesday. The round was led by Acrew Capital, with participation from Magnify Ventures, TTCER, and Resolute Ventures.

The inspiration for the company came from a horrifying incident involving the founders’ mother.

About two years ago, Patrick Coughlin’s mom called him, distraught, saying she had just received a phone call from a man saying he had kidnapped Coughlin’s sister. He was senior vice president of security products at Cisco at the time. (He landed there after Splunk bought his cloud security startup TruSTAR for a reported $82 million in May 2021. In 2024, Cisco bought Splunk.)

Her mobile phone rang with the caller ID of her daughter, Coughlin recounted. During that call, “she thinks she hears my sister’s voice saying, ‘Mom, they’ve got me.’ There’s a blood-curdling scream, and then my sister says, ‘You’ve got to do what they tell you.’ And then a man comes on the phone and says, ‘If you don’t pay us $1,200 right now, we’re going to kill your daughter in the parking lot of the local Walmart,’” he continued.

The scammer had accurately spoofed Coughlin’s sister’s number, her voice, and referenced the location of the Walmart she frequented.

Fortunately, the mom kept her wits, called the daughter, and discovered that she was fine. The kidnapping was an AI-generated scam.

Coughlin, like his mom, was shaken.

“What I was thinking, after calming my mom down is: What has fundamentally changed in the underlying cybercriminal economy that we are now able to lever the same kind of sophistication that I had seen pointed at government agencies, and then later at Fortune 500 companies? And now we’re deploying that sophistication at the consumer?”

The answer is, of course, cheap and powerful LLMs and other generative AI tools.

Before AI, pursuing such grifts on consumers was not financially worth it. It would require in-depth research on the target, tech to spoof voices, and so on. Such attacks were primarily aimed only at deep pockets, like enterprises or governments, as was the tech to defend against them.

“There’s something that’s happening right now to consumers with AI in the hands of cyber criminals,” Coughlin says. The costs to perpetrate such swindles have become negligible, and the research material, easily available.

“You can clone a voice off three seconds of audio, off a publicly available social media post. So we’ve all got these traces of stuff that’s out there in the ether — like where we’re talking or narrating; commenting on a kid’s football game while videotaping it, and putting it on Facebook.”

The FTC said last month that people reporting online crimes collectively lost $3.5 billion to imposter scams in 2025, triple the amount in 2020. While the majority of people reporting such scams are older Americans, some research implies Gen Z is also highly susceptible. Research from 2025 by Malwarebytes, a maker of antivirus and anti-malware tools, reported that Gen Z was targeted more often with text scams than other generations, and fell for them about 25% of the time.

The Coughlin brothers’ idea was to develop a real-time intervention tool.

They tested their idea, and the AI scam detection model they were building, by launching a free website called Scam Wise. It is anonymous, no registration required. Just upload any suspicious texts, photos, or emails, and Scam Wise will determine if it’s likely to be bogus.

“We launched that about four months ago. We’ve had 50,000 submissions, and it grows now every week by about 10,000 submissions or more,” Coughlin said.

Scam Wise proved a source of in-the-wild data to help train Savi’s scam-detection AI model. The startup is currently mostly using Google’s Gemini, but has built its software on an AI gateway, which allows it to tap other AI models as needed, like voice detection-specific options.

On Tuesday, Savi launched a paid product, an iOS and Android app for consumers, that can screen texts, voicemails, and incoming calls for scams.

Such features are available in a lot of different products (such as Malwarebytes), but Savi’s most impressive feature is live-call monitoring.

During a suspicious phone conversation, a user can opt to add the app’s live agent as a listener. Savi listens for behavioral tells that can identify if the situation is a grift while the call is in progress.

Savi’s fees are also a bit unusual. It charges $8/month, discounted to $63/year, to cover an entire family, and puts no cap on the number of users. So one plan can cover a person’s kids, spouse, parents, and that uncle who always seems to need tech support. Or whoever else that the primary account holder wants to add and provide administrative support to.

AI has changed the conditions for “how accessible being a fraudster is,” Coughlin said. “We’re creating fraudsters because we’re bringing down the barrier of deceiving people. So not only do we have the organized criminals and the syndicates behind this, but everyday people are sort of being tempted into playing fraud.”

Savi Security’s answer is like a new generation of anti-virus-like software: one that uses AI in real time just like the bad guys do.

*When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.*