The Duct Tape Marketing Podcast

Episode: "Why Voice AI Is Ready for Prime Time"

Host: John Jantsch
Guest: Ryan Murha, Founder of Yodel Fi
Date: February 25, 2026

Episode Overview

In this episode, John Jantsch discusses the rapid evolution and mainstream readiness of Voice AI with Ryan Murha, founder of Yodel Fi. The conversation explores how voice agents are transitioning from basic customer support bots to sophisticated roles within businesses—acting as sales qualifiers, schedulers, concierges, onboarding guides, and more. Ryan delves into the practical implementation, opportunities, ethical considerations, and future trajectory of voice-driven AI in business operations and customer experience.

Key Discussion Points & Insights

1. Defining Voice Agents and Their Evolution

• Voice Agents Explained:
  • Voice agents are built on large language models (LLMs) that are designed to produce natural, spoken responses, often using cloned voices of specific individuals (e.g., creators, brand ambassadors).
  • Moving beyond static chatbots or IVR (interactive voice response) systems, modern voice agents can adopt different persona “modes” matching various user interactions—support, sales, guidance, etc.
    • (01:14) "[A voice agent] is essentially just an LLM that knows it's supposed to respond in a way that's naturally speaking. And then you use another tool to have it actually read that text out loud as it's coming." — Ryan Murha

2. Personalization at Scale

• Advanced agents don’t just read answers—they act “in character,” adapting tone and conversational style to context (e.g., sales hat vs. customer service hat).

• Enabling loyal followers or customers to interact with brands or creators at a personal level, even as scale makes one-on-one human contact impossible.

  • (02:25) “We do also clone the voice... But... you as, say, a podcast host, you have a lot of different states... So what we try to do is we try to have a few different LLMs that an agent can call on and can be different versions of you and have different access to pieces of knowledge that you may need at a certain time.” — Ryan

3. Public Acceptance and Voice AI Maturity

• Growing mainstream comfort: As quality and reliability have improved, users recognize and accept interacting with AI—especially if it makes their lives easier.

• Prediction: 2026-2027 will be pivotal years for mass adoption of real, purpose-built voice agents.

  • (04:38) “I really think like 2026, 2027 are going to be the years of like real voice agents.” — Ryan

4. AI Agents as Critical Thinkers

• Yodel Fi uses “multiple levels” of LLMs; some orchestrate conversations while others perform targeted actions or call specialized agents for precise responses.

• Key mistake businesses make: expecting a single, generic AI model to replace complex roles instead of purpose-building agents for specific outcomes.

  • (05:40) “We use like multiple levels of LLMs... each LLM call is very targeted. And that's kind of the mistake that we're seeing a lot of businesses, like, fall into right now... they're using like a very general package.” — Ryan

5. Voice Agents in the Buyer’s Journey

• New consumer expectations: Users want deeper self-education before human contact; AI agents can answer detailed questions and support exploration without traditional sales pressure.

• Value creation: An AI that aids in decision-making is seen increasingly as a “value add,” not just a customer service tool.

  • (07:25) “People don’t want to be sold, you know. They just want to ask the questions... So in some ways like AI is kind of perfect for that.” — Ryan

6. Transparency, Ethics, and User Perception

• Some users may feel “duped” if they discover they were talking to AI; others appreciate the efficiency and convenience.

• Increasing trend toward upfront disclosure—labeling agents as “AI advisors” to foster trust and manage expectations.

  • (10:52) “People don’t want to feel that they're talking to an agent yet, but I do think that's going to change.” — Ryan
  • (11:17) “It should at least be on the website.” — Ryan

7. Implementation Frameworks: Purpose-Built Agents

• Start with focused, single use-case pilots; expand and layer complexity as comfort and ROI grow.

• Applications range from advanced facilitators to simple receptionists or schedulers.

• Strong emphasis on matching output to the brand's voice and customer experience goals.

  • (12:06) “We’re always going to start with a single small use case and try to nail that down and then kind of build things on top of it.” — Ryan

8. Guardrails & Hallucination Prevention

• Use of specific LLMs (e.g., Google Gemini) that are less prone to hallucinations.

• Guardrails in system prompts, robust testing, and restricting agents to verified knowledge bases or instructions.

• Hybrid/handoff models: “Press 0 for a human” remains an option.

  • (14:32) “I’m a big fan of the Gemini models... they definitely hallucinate less... if you give the agent the right context... it doesn’t have to go fill in the blanks itself.” — Ryan

9. Practical Deployment & Pilots

• Small agencies or creators can start with simple, single-prompt agents built on their own content.

• Customer feedback and iteration are key before scaling or rolling out multi-function agents.

• Options for using existing tools (e.g., ChatGPT custom GPTs) for quick, low-risk prototyping.

  • (18:22) “ChatGPT has these... custom GPTs... That’s a nice way to sort of test... Try it, see if it’s interesting for you, then you can engage us...” — Ryan
  • (19:09) “We’d do a single prompt, LLM... upload a few of your files and let you call it... Maybe show it to a few customers, get some feedback...” — Ryan

10. Industry Landscape & Yodel Fi’s Approach

• Yodel Fi consolidates best-in-class tech (e.g., Eleven Labs for voice cloning) and builds value-layered use cases.

• Focused not just on what’s possible, but on what’s valuable and enhances the customer or creator experience.

• Revenue opportunities for creators (integrate AI agent access into paid tiers).

  • (17:09) “We use 11 Labs. We're sort of a... combine the different pieces... and try to bring them to market... We're trying to meet people where it helps them.” — Ryan

Notable Quotes & Memorable Moments

• (03:38) “It's going to be... John in his sales hat and this is John in his customer service hat.”
• (08:23) “You can do [this] at three in the morning as well... it's instant.”
• (11:43) “I forgot to tell you when we booked this interview, I do need your Social Security number.” (John, jokingly)
• (13:34) “If you can offer me a product at a lower cost and because I speak to an AI agent and... Great, you know.”
• (17:41) “Some of the tools are like, well, okay, it's cool. It can do that. But like, why?”

Important Timestamps

| Timestamp | Segment | |-----------|--------------------------------------------------------------------------| | 01:14 | Defining voice agents versus LLMs and chatbots | | 02:25 | The importance of persona and voice cloning | | 04:38 | The state of public acceptance and maturity timeline prediction | | 05:40 | How agents “think” and common business mistakes in implementation | | 07:25 | AI’s emerging role as a value-add in the buyer journey | | 10:20 | User perception—feeling “duped” vs. appreciating AI convenience | | 11:17 | Ethics, transparency, and user notification | | 12:06 | Implementation process for voice agents (purpose-built framework) | | 14:09 | Hallucination prevention and hybrid AI-human models | | 17:09 | The tech stack: Yodel Fi’s approach, industry landscape | | 18:22 | How to safely pilot voice agents (advice for agencies, creators, brands) | | 19:09 | Smallest safe experiment for Yodel Fi |

Conclusion

Ryan Murha and John Jantsch’s discussion paints a clear picture: Voice AI is ready—and about to become essential—for delivering scalable, personal, efficient service and engagement. The path to success lies in specialized, purpose-built agents, careful matching with brand experience, and transparent customer communication. For businesses and creators, the opportunity to pilot accessible Voice AI is here, with real potential for customer delight and new revenue streams.

For more, connect with Ryan on LinkedIn or visit yodelfi.com for information and beta opportunities with Methotic.