Savi Security Launches AI-Powered Defense Against Realistic Scams

As generative AI lowers the barrier to entry for cybercriminals, a new startup is stepping in to provide much-needed consumer protection. Savi Security has officially launched its mobile application for iPhone and Android, aiming to thwart sophisticated AI-driven fraud like voice cloning and kidnapping rants.

The Rise of High-Sophistication Consumer Fraud

The impetus for Savi Security stems from a harrowing real-world incident where the founders' mother was targeted by a kidnapping scam. Utilizing AI, scammers successfully spoofed a family member's phone number, cloned her voice, and even referenced specific local landmarks to demand a $1,200 ransom.

Historically, the level of sophistication required to execute such "social engineering" attacks—involving deep research and high-end voice cloning—was reserved for targeting governments or Fortune 500 companies. However, the democratization of powerful Large Language Models (LLMs) has changed the economics of crime. Today, a fraudster can clone a human voice using just three seconds of audio scraped from a public social media post, making high-level deception cheap and scalable.

Real-Time AI Intervention and Detection

Savi Security, founded by tech veterans Patrick and Ryan Coughlin, addresses this threat by using AI to fight AI. The startup's technology is built on an AI gateway, which allows it to utilize various models—including Google’s Gemini—to detect specific fraudulent patterns.

One of the app's most significant technical differentiators is its live call monitoring feature. During a suspicious conversation, users can opt to have a live agent (supported by the app's detection models) listen in. The system monitors the call for behavioral "tells" that indicate a scam is in progress, providing real-time intervention to prevent the victim from falling for the grift.

To refine its detection capabilities, the team previously launched "Scam Wise," an anonymous tool that processed over 50,000 submissions. This provided a massive dataset of "in-the-wild" fraudulent texts, photos, and emails to train their proprietary models.

Scaling Protection for Families

The financial impact of these scams is staggering; the FTC reported that imposter scams cost victims $3.5 billion in 2025 alone. While older populations are frequently targeted, research from Malwarebytes suggests Gen Z is also highly vulnerable, falling for text-based scams at a rate of roughly 25%.

To combat this, Savi has adopted a family-centric business model. Instead of individual subscriptions, the service offers a flat fee of $8/month (or $63/year) that covers an unlimited number of users under one account. This allows tech-savvy users to provide a "security umbrella" for their children, elderly parents, and extended family members.

Key Takeaways

  • AI-Powered Defense: Savi uses an AI gateway to tap into various models, including Gemini, to detect voice cloning and sophisticated social engineering in real time.
  • Live Call Monitoring: The app offers a unique feature where AI-driven monitoring can listen to active phone calls to identify behavioral patterns typical of fraudsters.
  • Scalable Family Protection: The subscription model is designed for household security, allowing one account to protect an unlimited number of family members from text, email, and voice scams.