Regal AI launches Copilot for building self-improving voice AI agents  – News

Regal Voice Inc. today announced the launch of Copilot, which will allow companies to build and deploy artificial intelligence voice agents that self-improve over time without the hassle of prompting and engineering.

According to the company, the new copilot takes what would require days or weeks of engineering time and compresses it into hours and days, using already existing business logic. AI agents then learn and improve from customer conversations using best practices, fast, rational experimentation and flag underperformance to pivot quickly toward what works.

Regal’s roots are in AI contact center operations. The company says it fully understands the needs of customer service and where high-touch is needed for voice interaction. The company last raised $40 million in late 2024, bringing its total capital raised to $83 million.

The platform’s agents draw from millions of calls, Regal said, allowing the copilot to get teams started immediately. The copilot lowers the learning curve so that a business has a working voice agent within a day, allowing a team to construct guardrails, conversation flow and tune handoffs like pros. Since it has access to the business culture, data and anything else the team can throw at it, it can immediately incorporate brand style and tone up front and begin improving out of the box.

This allows teams to start with a simple voice agent if they choose to. Perhaps they don’t want something that does everything; instead, they just want a voice agent that does Frequently Asked Questions, handling delivery status, changing addresses, pausing and cancelling deliveries. 

Production teams can also use Copilot to modify the underlying system, add features and change how it operates. Regal explained that Copilot not only works with the team to assist with best practices, but it also asks clarifying questions, iterates on design to keep it within the framework and stress tests the design before deployment. 

It will even show its work, displaying reasoning so that the team can guide the build and get the final product they want faster. Once deployed, it analyzes the outcome of calls, including the sentiment and closure. It can also use AI-driven approaches to assist with new use cases such as outreach, expanded coverage and additional revenue generation opportunities such as personalization and regional trends.

Regal added that the Copilot will recommend fixes to the team based on what it learns as it operates, explaining its reasoning and why it will help from real call transcripts and “experience” of agents with human customers.

Image: News/Microsoft Designer