Tarini Padmanabhuni founded DetectifAI, a San Francisco startup, after her grandfather was duped by a deepfake voice scam impersonating his brother and demanding ransom. This incident, which occurred about two years ago, left a lasting impression on her because her grandfather had no reliable way to recognize the deception. With AI voice scams on the rise—Americans lost nearly $900 million to such fraud last year according to the FBI—DetectifAI aims to protect users by developing AI models small enough to run on smartphones, enabling real-time detection of fake voices without sending data to the cloud.

DetectifAI differentiates itself by creating compact AI models designed specifically to function directly within a phone’s operating system, offering instant analysis during calls or audio messages. Unlike competitors that rely on cloud-based detection services, this approach seeks to empower phone manufacturers to embed the technology natively, giving users built-in safeguards against voice deepfakes. Padmanabhuni envisions the partnership with phone makers as a strategic advantage similar to how AT&T’s exclusive deal affected the original iPhone’s market impact.

The core product is a software development kit (SDK) that phone makers and fraud prevention companies can license to integrate voice deepfake detection. DetectifAI already has some early revenue and handles over 100,000 calls monthly through AI voice agents in financial institutions across India, verifying speakers and detecting fraudulent voices during debt collection and loan follow-ups. Although the startup has not disclosed specific clients due to confidentiality, this traction highlights the practical application of its technology in high-risk sectors.

Backed by early seed investment from Josh Constine and Manohar Kamath, DetectifAI is gaining recognition and is set to compete in TechCrunch’s Startup Battlefield at Disrupt 2026 in San Francisco this October. Padmanabhuni, who brings a background in machine learning and cyber-physical systems, leads the company as it pursues licensing deals with handset makers and enterprises seeking to combat the growing problem of AI-enabled voice fraud.