
Jordan Belfort Launches AI Voice Startup Inspired by His Wolf of Wall Street Tactics
Author : Ravit Jain
Jordan Belfort, the convicted fraudster whose life inspired Martin Scorsese's blockbuster film, is now staking his claim in the artificial intelligence industry. The self-styled "Wolf of Wall Street" has co-founded a voice-based AI startup called Tone, built around his decades-old theories on sales psychology and the persuasive power of human tonality.
Belfort is a figure who has always managed to attract attention and, more remarkably, believers. His charisma is not difficult to understand, even for those who know his history well. In the same way that WeWork's Adam Neumann could draw investors into a vision that defied conventional financial logic, Belfort has a quality that makes people want to be in his orbit. That magnetism is now being channeled into a technology venture that sits at the crossroads of his most practiced skill - sales - and the most dominant trend in the technology industry today.
Belfort famously pleaded guilty in 1999 to securities fraud and money laundering connected to Stratton Oakmont, the Long Island brokerage he ran where salespeople used pump-and-dump schemes to push penny stocks on unsuspecting investors. During that period, Belfort accumulated significant personal wealth, including a Southampton estate, a 167-foot yacht, and a helicopter - two of which he destroyed during drug-fueled incidents later dramatized in Scorsese's film adaptation of his autobiography. He served 22 months in federal prison and was ordered to pay $110.4 million in restitution to 1,513 victims. Reports indicated that those victims ultimately received far less than the full amount owed, though Belfort reportedly completed his court-ordered restitution obligations in late April of this year.
Despite that history, Belfort has remained active in the public and commercial sphere. In recent years, he has operated as a Miami-based sales consultant and at one point hosted cryptocurrency workshops at his Miami Beach home, where attendance reportedly cost one bitcoin - or its cash equivalent of roughly $40,000 at the time.
Building a Voice AI Platform Around Persuasion Theory
Tone, Belfort's new venture, is designed to deploy AI-generated voices that do more than simply sound human. The platform's core premise is that the way something is said matters as much as what is actually said - a philosophy Belfort says he has been teaching since the late 1980s under the label "goal-oriented communication." The platform applies specific tonal patterns to AI-generated speech, such as framing declarative statements as questions to build engagement, or using a lower, near-whispered delivery to suggest scarcity of information.
Belfort says Tone is already being used by businesses that have more inbound sales leads than their human teams can realistically handle. Clients reportedly include law firms, health insurance companies, automobile dealers, and polling organizations. According to Belfort, the platform can conduct an entire health insurance sales conversation autonomously - asking qualifying questions, recommending a plan, explaining its features, adjusting the offer if a customer pushes back on price, and completing the transaction - all with a latency of approximately 600 milliseconds.
Now, the company is launching a self-service product aimed at content creators and authors who want to convert written material into AI-narrated audio. Users will be able to select from Tone's existing library of AI voices or, as a voice-cloning feature rolls out, generate a replica of their own voice. The move into content creation could also serve a secondary purpose: expanding the proprietary data used to train and refine the platform's models.
A Controversial Co-Founder and an Unverified Edge
Tone's co-founder is Shervin Pishevar, a longtime venture investor known for early bets on companies including Uber and Airbnb. Pishevar has faced misconduct allegations in the past, which he has denied. Neither Belfort nor Pishevar has an engineering background, which raises practical questions about the depth of their technical oversight. The company's patent lists a name separate from either co-founder, and Belfort identified Ben Sussman as the company's VP of engineering and the technical architect behind the product.
The two came together over dinner, Belfort explained, after he had been pitched by another AI company claiming its product could handle sales calls. He found that product unconvincing and began developing his own approach, eventually building the voice infrastructure from scratch after failing to find an existing voice technology that met his standards.
The AI voice market is already highly competitive, with well-funded players such as ElevenLabs and Cartesia offering sophisticated voice synthesis tools. Belfort acknowledges the competition but argues that the distinction Tone offers is not simply in how human its voices sound - he expects that gap to close across the industry within a year - but in the application of strategic tonality, which he says is protected by a patent.
Revenue, Funding, and Legal Compliance
Tone has generated approximately $700,000 to $1 million in revenue so far, according to Belfort, and has raised around $10 million through friends, family, and seed investors. Belfort says he plans to raise a more substantial round and does not see an immediate need to sell the company, though he acknowledged he would consider the right acquisition offer if one materialized.
On the regulatory side, Belfort said Tone operates strictly within the boundaries set by the Telephone Consumer Protection Act, which the Federal Communications Commission clarified in 2024 applies to AI-generated voices in the same way it applies to robocalls. Tone, he said, only contacts individuals who have opted in or who are existing customers. The exception is political and market polling, which carries its own statutory exemptions allowing calls to private residences without prior consent.
Regarding voice cloning and the risk of impersonation, Belfort said the company requires users to speak directly into the system to create a voice clone, generating a live voiceprint that is matched to the person giving consent. Uploaded audio files are discouraged and would require verification against a live voiceprint. He noted the cloning feature is technically ready but has been held back pending confidence in the safeguards.
Performance Benchmarks and the Role of Human Salespeople
Belfort was candid about the current limitations of Tone's AI relative to top human sales performers. He said a skilled human closer would still outperform the platform on a call-by-call basis, but expects that gap to narrow significantly within six to twelve months. He attributed the current ceiling partly to the decision to use smaller, faster proprietary models rather than large frontier models, which allows for low latency but comes at the cost of raw intelligence.
He was also clear that the platform is not designed to eliminate human sales staff. In his view, the primary use case is coverage - handling volume that exceeds what a human team can manage, operating across time zones, or engaging in languages that a given team may not support. For high-value transactions where customers expect and demand human interaction, he believes skilled salespeople will remain essential.
One of the more specific technical claims Belfort made concerns how the AI handles conversations that deviate from an ideal script. He said the system is trained to navigate off-topic questions and objections, respond naturally without hollow affirmations, and return to the core sales narrative without losing conversational rapport or producing inaccurate information - what he characterized as the AI equivalent of a salesperson lying.
Addressing the Trust Problem Directly
When asked directly how potential enterprise clients - banks, hospitals, consumer brands - should reconcile their due diligence with the fraud conviction of one co-founder and the misconduct allegations against the other, Belfort offered a straightforward response. He acknowledged the mistakes of his past, pointed to more than two decades of work as a sales trainer and consultant hired by companies around the world, and said that businesses unwilling to work with him were entitled to that decision. He expressed confidence that enough companies would see value in the product regardless.
The pricing model for Tone varies by use case. Streaming voice services are billed per minute. The content creator platform will operate on a monthly subscription basis. Some enterprise clients are structured around performance-based compensation, where Tone earns a larger share of revenue tied to successful outcomes.
Belfort also addressed a Paramount+ documentary released this past summer that revisited his story. He said he has not watched it, adding that he makes a deliberate practice of avoiding coverage about himself - positive or negative - and relies on his own internal sense of who he is rather than external validation or criticism.
Whether Tone ultimately delivers on its promise remains to be seen. The voice AI space is crowded, the technology's claimed advantages have not been independently verified, and the backgrounds of its founders will invite scrutiny from any serious institutional buyer. But the history of technology investment is filled with ventures that attracted capital and customers on the strength of a compelling narrative and a founder who knew how to tell it. Belfort, whatever else might be said about him, has never struggled with that part.
Source Credits : TechCrunch


