Savi’s app aims to protect consumers from realistic AI scams like kidnappers demanding ransom

Savi's app aims to protect consumers from realistic AI scams like kidnappers demanding ransom

The consumer security market has long operated on a flawed assumption: that sophisticated attacks only target enterprises with deep pockets. Savi Security is betting that assumption just shattered.

The startup, founded by brothers Patrick and Ryan Coughlin, launches its iPhone and Android app Tuesday backed by $7 million in seed funding led by Acrew Capital, with Magnify Ventures, TTCER, and Resolute Ventures participating. Their thesis is direct: generative AI has collapsed the cost structure of targeted social engineering, bringing nation-state-grade deception to anyone with a smartphone and a few dollars in API credits.

The incident that catalyzed a company

The origin story reads like a thriller. Two years ago, Patrick Coughlin — then SVP of security products at Cisco after Splunk acquired his cloud security startup TruSTAR for a reported $82 million — received a frantic call from his mother. A man claiming to have kidnapped Coughlin's sister demanded $1,200. The caller ID displayed his sister's number. His mother heard what sounded like her daughter's voice: "Mom, they've got me," followed by a scream, then "You've got to do what they tell you."

The scammer knew the Walmart location the sister frequented. The voice clone was convincing enough to shake a career security executive's mother. She kept her wits, called her daughter directly, and confirmed the kidnapping was fabricated. But the psychological impact lingered.

"What has fundamentally changed in the underlying cybercriminal economy," Coughlin recalled thinking, "that we are now able to lever the same kind of sophistication that I had seen pointed at government agencies, and then later at Fortune 500 companies? And now we're deploying that sophistication at the consumer?"

Economics of the AI scam supply chain

The answer sits at the intersection of three commoditized capabilities: voice cloning from three seconds of public audio, caller ID spoofing infrastructure available on grey markets, and LLMs that can orchestrate real-time social engineering at near-zero marginal cost. Pre-AI, this level of personalization required human researchers and custom tooling — viable only for high-value targets. Today, a criminal can automate reconnaissance across social media, generate a voice model, and execute thousands of tailored calls for pennies.

Consumer defenses haven't kept pace. Carrier-level STIR/SHAKEN attestation remains spotty. Email filters catch phishing but not vishing. Password managers and 2FA protect credentials but not human judgment under duress. The industry has effectively left households unarmed against attacks that mimic trusted contacts in real time.

Savi's architectural bet

Savi's app operates as a consumer-grade security operations center. It ingests communications across voice, SMS, and email, applying behavioral analysis and threat intelligence to flag anomalies — spoofed numbers, synthetic audio artifacts, urgency tactics, and known scam patterns. The goal isn't just detection; it's interruption. When the app identifies a probable AI-generated scam call, it can block, warn, or route the user to a verification workflow before money moves.

The Coughlins' backgrounds inform the approach. Patrick's TruSTAR and Cisco tenure focused on threat intelligence sharing and enterprise detection pipelines. Ryan's consumer product experience at Apple and Spotify emphasizes frictionless UX — critical for an app that must earn daily permission to monitor communications without inducing alert fatigue.

A category in formation

Savi enters a nascent category. Truecaller and Hiya dominate caller ID and spam blocking but lack real-time AI audio analysis. Enterprise-oriented platforms like Pindrop and Nuance (now Microsoft) focus on contact center fraud prevention, not consumer endpoints. Startups such as Aura and IdentityForce bundle identity monitoring with insurance — reactive rather than preventive.

The $7 million seed round signals investor conviction that consumer security is due for a platform shift comparable to the endpoint protection wave of the 2010s. Acrew Capital's leadership suggests a thesis around network effects: as more users adopt Savi, crowdsourced threat intelligence improves detection for everyone, creating a data moat.

Open questions

Execution risks are substantial. Apple's iOS restrictions on call screening and SMS access limit what any third-party app can intercept. Android offers more flexibility but fragments across OEMs. Savi will need clever workarounds — possibly leveraging call forwarding, VPN profiles, or carrier partnerships — to achieve comprehensive coverage.

Privacy scrutiny will intensify. An app that analyzes voice calls and messages in real time sits on a fulcrum of trust. Transparent architecture, on-device processing where feasible, and clear data retention policies will determine whether security-conscious users adopt or avoid.

Monetization remains undefined. The launch is free; a subscription tier seems inevitable. Consumer willingness to pay for proactive scam prevention — versus reactive identity theft insurance — is unproven at scale.

The broader implication

Savi's launch matters beyond its own traction. It marks the moment the security industry acknowledges that the AI threat landscape has fully democratized. The same tooling that enables a $1,200 kidnapping scam today will enable a $1.2 million business email compromise tomorrow, and the defensive lessons transfer.

For RevOps and security leaders watching from the B2B side, the signal is clear: the boundary between consumer and enterprise threat surfaces has dissolved. Employees bring personal devices, personal numbers, and personal vulnerabilities into corporate environments. A security strategy that ignores the consumer attack surface is incomplete.

Savi may succeed or fail on product execution. But the problem it targets isn't going away. As voice cloning costs trend toward zero and LLM reasoning improves, the only sustainable defense is automated, intelligent interception at the point of contact — for everyone.