Savi: App to Protect Against AI Voice and Call Scams

Photo: TechCrunch
Quick answer
Savi Security has developed an AI-powered app to shield users from AI-driven voice and call scams, including deepfake impersonations.
Savi Security, founded by former Cisco, Apple, and Spotify employees, has launched an app designed to combat AI-driven fraud. The solution detects fake voices, calls, and messages that are becoming increasingly realistic due to advancements in generative AI models.
The idea for Savi emerged after the founders' mother received a call from "kidnappers" using a cloned voice of their sister. The scammers used a spoofed number and personal details to convince the victim of the threat. Though the deception was uncovered, the incident highlighted how accessible tools for large-scale fraud have become.
The Savi app leverages AI to analyze incoming calls, messages, and voice notes in real time. A key feature allows it to join live conversations to assess their authenticity. The service has already secured $7 million in funding from Acrew Capital and other investors. Subscriptions cost $8 per month or $63 per year for unlimited family use.
According to the U.S. Federal Trade Commission, losses from identity fraud reached $3.5 billion in 2025—three times higher than in 2020. While older adults are often targeted, research shows Gen Z is also vulnerable, particularly to text-based scams.
Savi Security positions its solution as a next-gen antivirus, using the same AI technologies as scammers but for protection. The company also offers a free Scam Wise service to verify suspicious messages, which has already processed over 50,000 requests.
Common questions
- How does the Savi app work?
- Savi uses AI to analyze incoming calls, messages, and voice notes in real time, detecting signs of fraud such as cloned voices or suspicious behavioral patterns. The app can also join live conversations for additional verification.
- Why has AI scamming become so widespread?
- The availability of advanced language models and voice cloning tools has made fraud easier and cheaper to execute. Scammers can now clone voices using just a few seconds of audio from social media, making attacks more realistic and scalable.
- Who is most vulnerable to AI scammers?
- While elderly individuals are frequent targets, Gen Z is also at risk, particularly from text-based scams. In 2025, losses from such attacks exceeded $3.5 billion!
Dzen feed: /feed/dzen.xml · RSS: /feed.xml