AI scams are becoming more personal, more convincing and harder to dismiss in the moment. Savi Security, a new startup from brothers Patrick and Ryan Coughlin, is entering that problem with an app meant to help everyday people recognize fraud across texts, emails, voicemails and phone calls.
The company launched its app for iPhone and Android on Tuesday after raising $7 million in seed funding. The round was led by Acrew Capital, with participation from Magnify Ventures, TTCER, and Resolute Ventures.
A family scare became the company’s starting point
The idea behind Savi Security came from an incident involving the founders’ mother about two years ago. Patrick Coughlin’s mom received a call that appeared to come from her daughter’s phone number. A man claimed he had kidnapped Coughlin’s sister.
According to Coughlin’s account, the call included what sounded like his sister’s voice, a scream, and a demand for $1,200. The caller also referred to a local Walmart that the sister frequented. The scammer had accurately spoofed the sister’s number, her voice and a familiar location.
The kidnapping was not real. Coughlin’s mother called her daughter and learned she was safe. But the episode showed how a scam could combine caller ID spoofing, voice imitation and personal context into a single high-pressure moment.
At the time, Patrick Coughlin was senior vice president of security products at Cisco. He had arrived there after Splunk bought his cloud security startup TruSTAR for a reported $82 million in May 2021. In 2024, Cisco bought Splunk. Ryan Coughlin brought consumer product experience from Apple and Spotify.
Why AI changes the economics of fraud
Savi Security is built around a simple concern: scams that once required serious effort can now be made more cheaply and convincingly. Before generative AI tools became widely available, an attacker trying to deceive a specific person needed research, technical tools and enough potential payoff to justify the work.
That kind of sophistication was more commonly aimed at enterprises, governments and other high-value targets. The defenses were also usually built for those organizations, not for ordinary consumers trying to decide whether to answer a call or trust a message.
Patrick Coughlin argues that cheap and powerful LLMs and other generative AI tools have changed that balance. Research material is easier to find, and voice cloning can be built from small public traces of audio, such as social media videos.
The broader fraud problem is already large. The FTC said last month that people reporting online crimes collectively lost $3.5 billion to imposter scams in 2025, triple the amount in 2020. The source article also notes that older Americans make up the majority of people reporting such scams, while research from 2025 by Malwarebytes reported that Gen Z was targeted more often with text scams than other generations and fell for them about 25% of the time.
From Scam Wise to a paid app
Before launching the consumer app, the Coughlin brothers tested their approach with a free website called Scam Wise. The site is anonymous and requires no registration. Users can upload suspicious texts, photos or emails, and the service evaluates whether the material is likely to be fake.
Coughlin said Scam Wise launched about four months ago and has received 50,000 submissions. He also said it now grows every week by about 10,000 submissions or more.
Those submissions gave Savi access to in-the-wild examples that helped train its scam-detection AI model. The startup is currently mostly using Google’s Gemini, but its software is built on an AI gateway. That setup allows Savi to use other AI models as needed, including options built for voice detection.
The paid Savi app extends that idea from uploaded examples to active consumer protection. It can screen texts, voicemails and incoming calls for scams. Similar features exist in other products, including Malwarebytes, but Savi is emphasizing live-call monitoring as its most notable feature.
Live-call monitoring is the central bet
During a suspicious call, a user can choose to add Savi’s live agent as a listener. The app then listens for behavioral signals that may indicate the conversation is a scam while the call is still happening.
That timing matters. Many imposter scams rely on urgency, fear and confusion. A person may not have time to investigate carefully if the caller is demanding immediate action or claiming that a family member is in danger.
Savi’s approach is to put the detection tool into the moment when the pressure is highest. Instead of asking consumers to analyze the situation alone afterward, the app is meant to intervene while the conversation is unfolding.
The company is also pricing the service around families rather than individual seats. Savi charges $8/month, discounted to $63/year, for an entire family and does not cap the number of users. One plan can cover children, a spouse, parents or other people the primary account holder wants to add and help administer.
A consumer version of security software for the AI era
Savi Security’s pitch is that consumer protection needs to evolve because the tools used by scammers have changed. AI has lowered the barrier to deception, making realistic fraud attempts easier to create and aim at ordinary people.
That does not mean every suspicious message or call is new. Text scams, imposter scams and spoofed calls already existed. The difference Savi is focused on is the quality and personalization that generative AI can bring to those attacks.
The company’s answer resembles a new generation of anti-virus-like software. It uses AI in real time to identify AI-enabled fraud, with a focus on the family members and everyday situations that traditional enterprise security products were not built to protect.