Interview with Alex Levin, Co-Founder and CEO of Regal, high-ranking executive featured in a multi-part discussion series.
Regal, a voice AI platform for businesses, was launched in 2020 by Alex Levin, Co-Founder and CEO. Levin, dissatisfied with traditional contact center software vendors' focus on "deflection" and "automation" over customer needs, saw the potential in voice as the most important channel and the ability to build orchestration, A/B testing, and personalization tools[1].
Regal's features include a no-code builder, real-time analytics, A/B testing, and built-in compliance for regulated industries. The platform integrates with CRM systems and can handle millions of conversations at scale. It provides AI voice agents for sales, support, scheduling, and collections[1].
One of the key technical challenges Regal faced was achieving low latency in voice AI agents for live interactions. To address this, Regal uses Retrieval-Augmented Generation (RAG) to improve performance without sacrificing responsiveness or natural cadence[1]. This approach lowers latency and allows the AI agent to keep talking with the customer if retrieval takes longer, maintaining engagement and a human-like conversational pace.
Creating high-fidelity voice replicas of professional voice actors and investors requires only 5-10 minutes of high-quality audio. However, ethically, consent from the original voice owners is crucial to prevent misuse. Regal recommends the use of unique family "safe words" to authenticate the real person behind the voice if cloned[1].
Regal's system for personalizing conversations uses a unified customer profile, LLMs, and Custom Actions that can leverage other services and ML. The company initially focused on tools for contact centers to manage human agents more effectively.
Regal's AI advisors for founders have been well-received and are used for various purposes, including product roadmap and pricing model advice. The platform is expected to evolve to perform better than human agents, with potential improvements in autonomous decision-making, emotional intelligence, and multimodal support.
Regal uses A/B testing to measure and improve model performance over time. The company is also working on improving automated evaluations to cut out manual QA for every change to every AI agent. Reinforcement learning is not currently used due to legal concerns, but may be introduced in 13 months.
Integrating Regal with major CRMs, payment systems, and internal APIs can be straightforward, but access to necessary APIs is a challenge. The investor agents are trained on the individual VC's prolific writings and philosophies.
In conclusion, Regal is a revolutionary voice AI platform that is transforming the way businesses interact with their customers. By focusing on natural conversations, personalization, and low latency, Regal is setting a new standard for voice AI in contact centers and beyond. To learn more about Regal, visit their website.
[1] Source: Regal's official website and press releases.
The AI voice agents provided by Regal can be utilized not only in sales, support, scheduling, and collections but also in investing, as the technology can clone the voices of professional voice actors and investors, given the necessary consent.
In the future, Regal's AI advisors for founders are anticipated to surpass human capabilities, offering improvements in areas like autonomous decision-making, emotional intelligence, and multimodal support, as the platform continues to evolve.
