To Their Credit

Voice and audio

Tone launches AI voice platform built to steer delivery toward a goal

Straightline.AI’s system combines conversational sales agents, expressive narration and consent-based voice cloning.

Ledger: AdoptedAI is taken up

Straightline.AI has launched Tone, a voice-based AI platform that alters vocal delivery as a conversation unfolds, choosing tone, volume, speed, pacing and emphasis according to the audience, context and intended outcome. Announced in New York on Nov. 16, Tone was co-founded by Jordan Belfort and Shervin Pishevar and is aimed at businesses, developers and creators.

The company’s central claim is that the system determines not only what to say but how to say it for a particular communication goal. It draws on Belfort’s Straight Line Tonality methodology, which examines how vocal expression affects how a message is understood. “Two people can say exactly the same thing and produce completely different reactions,” Belfort said. “The difference is often tonality.”

Tone combines real-time conversational agents, expressive speech generation and voice cloning. Straightline.AI said the technology is supported by U.S. Patent No. 12,437,153 B1, “Persuasive conversational agents for selecting words and tonalities,” granted in October 2025. The patent covers selecting language and vocal delivery according to communication objectives.

For business users, the company is positioning conversational agents for customer service, sales, education, hospitality and digital assistance. The agents include persistent memory, knowledge-base integration, workflow automation and connectivity to enterprise systems, as well as objection handling and active listening. Belfort told TechCrunch that the system has roughly 600 milliseconds of latency and uses smaller, lightweight models to respond quickly.

Belfort said customers include law firms, health-insurance companies, polling operations and automobile-related businesses. He described an insurance workflow that asks qualifying questions, selects and explains a plan, then changes the recommendation if a customer rejects it. He said companies use the system when they cannot contact all their leads with existing sales teams, or need coverage across languages and times of day.

Tone does not yet match a top sales producer, Belfort said, though he argued it could outperform a mediocre closer. He estimated its revenue at about $700,000 to $1 million and said it had raised about $10 million in friends-and-family and seed funding. Its pricing includes per-minute streaming charges, monthly subscriptions to the content platform and success-based fees for some clients.

The product also offers a self-service platform for creators and authors turning books into expressive AI narration. Its speech-generation tools are intended for audiobooks, podcasts, advertising, educational material, digital media and brand communications. Users can choose from a library of AI voices, while Tone is rolling out a way for users to clone their own voice; English and Spanish capabilities were introduced at launch, with further voices and features planned.

Straightline.AI said authorized users and businesses can create voice clones. Belfort told TechCrunch that creating a clone requires a person to speak into a computer and sign consent, while an uploaded recording must be matched to a live voiceprint in real time. He also said outbound calls go only to opted-in leads or existing customers, apart from polling, which he described as subject to an exemption for calls to private homes without opt-in.

The platform’s voice features operate in a regulatory environment that includes rules on synthetic speech. A research report on global voice-cloning regulation notes that AI-generated robocall voices in the United States are covered by existing consent requirements and penalties under the Telephone Consumer Protection Act following an FCC ruling in February 2024. The European Union’s AI Act transparency rules for qualifying manipulated or generated audio have applied since Aug. 2, 2026, requiring disclosure or labeling.

Tone was available in research preview at launch, with voice samples, a voice library, interactive demonstrations of goal-oriented communication and conversational-agent tests. The company is offering tailored agent deployments and API access as it pursues its stated goal of making goal-oriented voice intelligence a foundational AI capability.

More from Voice and audio

All Voice and audio news
  1. Equity ballot puts UK film and TV AI consent rules to the testLedger: Adopted
  2. Murder Drones actor says Voicemod used her voice without consentLedger: Disputed
  3. SAG-AFTRA chief says rules, not an AI ban, are the union’s pathLedger: Adopted