Savi raises $7M to shield families from AI voice-clone scams
A Cisco veteran's startup wants to put an AI bodyguard on every phone after his own mother nearly wired ransom to a cloned voice. Can consumers finally fight back?
The News
Savi Security, a consumer-safety startup founded by brothers Patrick and Ryan Coughlin, has closed a $7 million seed round to take on a fast-growing menace: scams supercharged by artificial intelligence. The round was led by Acrew Capital, with participation from Magnify Ventures, TTCER and Resolute Ventures.
The company is launching its app on iOS and Android on 8 July 2026. Rather than simply blocking spam, Savi screens incoming texts, voicemails and calls for signs of fraud, and offers a live-call monitoring feature in which a Savi agent listens in on a suspicious call and flags manipulation in real time. The technology is built on Google's Gemini, routed through a gateway that lets the team swap in other models.
Pricing is set at $8 a month or $63 a year, and a single subscription covers an entire family with no cap on the number of users. The founders bring heavyweight resumes: Patrick Coughlin was previously senior vice president of security products at Cisco, having earlier worked at Splunk and cloud-security startup TruSTAR, while Ryan Coughlin held consumer-product roles at Apple and Spotify.
The idea was deeply personal. Patrick's mother received an AI voice-clone kidnapping call that spoofed his sister's number and voice, demanding a $1,200 ransom.
Why It Matters
The scale of the problem is no longer theoretical. The US Federal Trade Commission logged $3.5 billion in losses to imposter scams in 2025, roughly triple the figure from 2020. Research from Malwarebytes in 2025 found that Gen Z is increasingly targeted by text-based fraud and falls for it about a quarter of the time.
What makes Savi notable is the inversion it represents. For a decade, enterprise-grade security tooling stayed locked inside corporate networks while ordinary consumers made do with crude spam filters. "What has fundamentally changed in the underlying cybercriminal economy that we are now able to lever the same kind of sophistication," Patrick Coughlin asked, framing the bet that the same AI arming fraudsters can be turned back on them.
The last comparable shift was the antivirus boom of the late 1990s, when McAfee and Norton turned a niche IT concern into a household purchase. Savi is wagering that voice-cloning fraud will do the same for AI-era defence.
Indian Angle
Nowhere is this threat more acute than India. The government's Indian Cyber Crime Coordination Centre has repeatedly warned about "digital arrest" scams, in which callers impersonate police or CBI officers and pressure victims into transferring money, a racket the Prime Minister himself flagged on national radio. Voice-cloning of a relative in distress is now a staple of the Indian fraud playbook, and citizens lost thousands of crores to cyber fraud in the past year alone.
Yet the Indian consumer has few purpose-built defences. Truecaller dominates call identification here, but it flags spam rather than dissecting the content of a live scam call. A Savi-style product priced for Indian wallets, or a domestic equivalent from a startup in the Razorpay or CRED orbit, could find enormous demand, especially as UPI makes instant, irreversible transfers the default.
Regulators are circling too. The RBI has tightened rules on mule accounts, TRAI has cracked down on spoofed caller IDs, and MeitY is weighing deepfake labelling norms. An AI listener that intercepts fraud before the money moves would slot neatly into that agenda, though data-privacy questions under the DPDP Act would need careful answers before Indian users hand a startup access to their calls.
FAQ
When does the Savi app launch?
Savi becomes available on both iOS and Android on 8 July 2026. The company's free Scam Wise website has been live for around four months and has collected over 50,000 scam reports that help train its detection systems.
How much does it cost?
The subscription is $8 per month or $63 per year. Unusually, one plan covers a whole family with no limit on the number of members, a pricing choice aimed at protecting elderly relatives who are common scam targets.
How is this different from Truecaller in India?
Truecaller mainly identifies and blocks unknown or spam numbers. Savi goes further by analysing the substance of texts, voicemails and even live calls for signs of manipulation, using Google's Gemini to judge whether a caller is running a scam script.
Where can I read the original report?
The details were first reported by TechCrunch, whose coverage is linked below.
This story was reported by TechCrunch. Read the full original coverage at TechCrunch.