Patrick and Ryan Coughlin have launched Savi Security. The brothers, formerly of Cisco, Splunk, Apple, and Spotify, are betting on a new wave of security apps. Their goal is to stop realistic AI scams that flood text, email, and phone lines.
The company raised $7 million in seed funding from Acrew Capital, Magnify Ventures, TTCER, and Resolute Ventures. The app goes live on iPhone and Android on Tuesday.
The family call
The idea started with a call to the Coughlin brothers’ mother. Two years ago, she received a message from a man claiming her daughter was held captive. He demanded $1,200 immediately or he would kill the girl in a Walmart parking lot.
The caller ID showed her daughter’s number. Her phone rang with what sounded like her voice screaming, “Mom, they’ve got me,” followed by the instruction to obey the kidnapper. The scammer had spoofed the number and voice, and correctly named the Walmart her daughter frequented.
The mother checked on her daughter, who was safe. The kidnapping was an AI-generated hoax.
Patrick Coughlin was shaken by the event. He noted a shift in the cybercriminal economy. Sophistication once reserved for government agencies and large corporations is now being used against ordinary people.
Before AI, these attacks were too costly for criminals. They required deep research and complex voice spoofing tech. Now, the tools are cheap and powerful. Large language models and generative AI tools make it easy to clone voices from three seconds of audio found on social media.
People leave traces everywhere. Narrating a football game or commenting on a video creates data that criminals can use.
The Federal Trade Commission reported that people lost $3.5 billion to imposter scams in 2025. That is triple the amount from 2020. While older Americans report most incidents, research from Malwarebytes suggests Gen Z is also highly vulnerable. In 2025, Gen Z fell for text-based scams about 25% of the time.
How it works
The Coughlin brothers built a real-time intervention tool. They tested it with a free website called Scam Wise. Users upload suspicious texts, photos, or emails without registering. The site determines if the content is likely fake.
Four months ago, Scam Wise received 50,000 submissions. That number grows by about 10,000 every week. This data trains Savi’s detection model.
The startup currently uses Google’s Gemini but runs on an AI gateway. This lets them tap into other models, such as those specialising in voice detection, as needed.
Savi’s paid app screens texts, voicemails, and incoming calls. While other products like Malwarebytes offer similar features, Savi includes live call monitoring.
During a suspicious conversation, a user can add the app’s live agent as a listener. Savi listens for behavioural tells to identify a scam while the call is in progress.
The pricing covers an entire family for $8 a month, or $63 a year. There is no cap on the number of users. One plan can cover kids, spouses, parents, and other relatives needing tech support.
Coughlin explained that AI lowers the barrier for deceiving people. This means not only organised syndicates, but everyday individuals, are tempted to commit fraud.
Savi Security acts like a new generation of anti-virus software. It uses AI in real time, just as the bad guys do.
