A Fake Kidnapping Call Inspired a $7M Startup to Fight AI Voice Scams

Patrick Coughlin's mother picked up the phone about two years ago and heard her daughter's voice screaming that she had been kidnapped. The caller ID matched her daughter's number. The voice sounded exactly right, down to a specific detail about the Walmart she regularly visited. It was all fabricated. A quick call to the daughter confirmed she was safe, and the "kidnapping" turned out to be an AI-generated impersonation designed to extract a ransom over the phone.

That incident became the founding story behind Savi Security, the company Coughlin built with his brother Ryan to help consumers detect exactly this kind of fraud. On Tuesday, July 7, 2026, Savi launched its app for iPhone and Android, arriving with $7 million in seed funding led by Acrew Capital. Magnify Ventures, TTCER, and Resolute Ventures also participated in the round.

The app screens incoming calls, texts, and voicemails for signs of AI-driven scams, the same category of fraud that nearly convinced Coughlin's mother a family member had been abducted. Rather than relying solely on blocklists of known scam numbers, the tool is built to catch cloned voices and spoofed caller IDs, the specific tactics that make modern impersonation attempts far more convincing than the crude robocalls of a decade ago.

Coughlin's background is squarely in cybersecurity. He previously worked in national cyber defense before founding TruSTAR, a cloud security startup that Splunk acquired for a reported $82 million in May 2021. Splunk itself was later acquired by Cisco in 2024, giving Coughlin a firsthand view of enterprise-grade threat detection before turning his attention to consumer protection. His brother Ryan brings a different skill set, having worked on consumer products at Apple and Spotify, experience that shaped Savi's approach to making a security tool approachable for everyday users rather than IT departments.

A Fraud Problem Already Measured in Billions

The timing reflects a fraud landscape that has scaled quickly. The FTC reported that imposter scams cost consumers $3.5 billion in 2025 alone, a figure that includes the kind of voice-cloning and fake-emergency schemes that targeted the Coughlin family. Generative AI has made it trivial to reproduce a person's voice from short audio samples scraped from social media or voicemail greetings, lowering the barrier for scammers to run highly personalized cons at scale.

Savi says its app had already processed 100,000 user submissions as of launch day, an early adoption signal for a product that had not yet been publicly available. Whether that volume reflects genuine demand for AI-scam detection or a wave of curiosity-driven downloads will become clearer as usage data accumulates over the coming months. Still, the underlying problem the Coughlin brothers are targeting is well documented: as synthetic voice and video become nearly indistinguishable from the real thing, the tools consumers have traditionally relied on to sniff out fraud, gut instinct and caller ID, are no longer reliable defenses on their own.