After a deepfake voice fooled her grandfather, this founder sprang into action

What happened

San Francisco-based startup DetectifAI was founded by Tarini Padmanabhuni, prompted by her experience of her grandfather being scammed into paying ransom via a deepfake voice fraud two years ago. The company rejected the common industry approach of downscaling cloud-hosted large models to fit mobile devices, instead building lightweight models designed from the ground up to be directly embedded into mobile operating systems. This enables real-time deepfake detection for calls and voice messages, with the entire detection process running locally so no audio leaves the user’s device. The platform currently processes over 100,000 AI voice call detection requests per month for Indian financial institutions, with future plans to expand licensing services for anti-fraud organizations.

Key facts

Company name

DetectifAI (headquartered in San Francisco)

Core product

On-device deepfake voice detection SDK embeddable into mobile operating systems

Current business scale

Processes over 100,000 call detection requests per month for Indian financial institutions

Industry loss data

U.S. losses from AI-powered scams reached nearly $900 million last year, up 24% from 2024

Funding update

Completed a small seed funding round, with investors Josh Constine and Manohar Kamath

Background

AI-powered voice fraud is growing rapidly in scale, with losses incurred by adults over 60 reaching twice the amount recorded for the 50-59 age group. Most existing mainstream deepfake detection products are deployed on cloud servers, so mobile manufacturers cannot natively integrate them into their devices, leaving average users without effective protection for real-time call scenarios.

Why it matters

If this technology is natively built into mobile devices, average users will be able to identify AI-generated fake voices in real time during calls. Since no user audio is transmitted during the detection process, the solution balances robust privacy protection with strong fraud defense capabilities. For mobile manufacturers, being the first to launch with this feature will create a differentiated competitive edge, while filling a critical technical gap in the deepfake detection ecosystem for on-device use cases.

In their words

“What stuck with me was not the money my grandfather lost to the scam, but that he had no way to tell at the time that the voice he heard was AI-generated.”

DetectifAI founder Tarini Padmanabhuni

“Just as AT&T’s exclusive partnership with the original iPhone launch helped carriers pull ahead in competition, the first mobile manufacturer to integrate our detection feature will gain significant competitive advantage.”

DetectifAI founder Tarini Padmanabhuni

What to watch

Stakeholders can look forward to DetectifAI’s presentation at the October TechCrunch Disrupt startup competition, as well as updates on its embedded integration partnerships with mobile device makers.