Regal AI Voice Agents: Live Demo & Performance Review

Subscribe to our channel

Regal AI offers tools to build, test, and optimize voice agents for automating customer conversations. But how well do those agents handle unexpected questions, interruptions, and requests beyond their capabilities?

In this live review, Katherine Stone puts two Regal AI voice agents to the test: a David Frankel AI persona and a Shih Tzu hotline. She evaluates their conversational abilities, guardrails, voice quality, context retention, and responses to challenging scenarios.

What We Tested

• Pricing questions and objection handling
• Guardrails and requests for unauthorized discounts
• SMS, multilingual support, and human transfer requests
• Voice quality, latency, pacing, and expressiveness
• Interruption handling and conversational flow
• Context retention throughout the conversations

The first agent handled objections and rejected a request for free access, but couldn't send an SMS, switch languages, or transfer to a human. The Shih Tzu agent demonstrated good context retention and expressive speech, though its pacing and interruption handling needed improvement.

Watch the full review to see both agents in action, how they respond to real-world scenarios, and where their performance falls short.