Kevin: In today’s episode, I speak with Michael LeVene. Michael shares fantastic insights on how we can use voice AI to improve our communication with customers — particularly by making a massive return on investment through reception staff. The ability of AI to never miss a call, answer client and prospect questions, and book appointments is phenomenal. You’re also going to get two great demonstrations during today’s session.
I can’t wait. Let’s jump in. Michael, welcome to the call today.
Michael: Thank you, Kevin. I’m very excited about this topic. We’re diving into the field of AI today. While many people are experimenting with tools like ChatGPT, image generation, and video AI, I think a lot of people underestimate the value of audio in AI. I know you’re an expert in that area, so I’d love to hear your insights about what’s exciting in this space and what we should be using right now.
Michael: Thanks. One of the things I really like about AI is that it can now do work computers couldn’t do well before — especially voice AI. It understands you and can respond in a humanlike way, which opens up huge opportunities for businesses.
AI can answer phones, handle random conversations, book appointments, answer questions, and even quote customers if the information is available. It can essentially replace or support receptionists, freeing up staff for more in-depth, productive work — especially in businesses with repetitive inquiries.
One area I focus on is medical practices, which receive thousands of phone calls per month — mostly for booking, shifting, or cancelling appointments. That’s where we’re seeing very clear paybacks.
I could give an example — perhaps bring up one of my assistants. In fact, the one that picks up if I don’t answer my phone. You can have a chat with it to see what an AI can do on the phone.
Kevin: I’d love that, because my question was going to be, “Does this work well?” So instead of asking, let’s demo it.
Let’s give it a call. This is live, not pre-set.
AI Assistant: Kevin, it seems Michael is busy. What message would you like to leave?
Kevin: Hey, I was wondering if Michael might be able to give me a demo of how these AI tools work.
AI Assistant: That sounds great, Kevin. Michael would be happy to give you a demo. Would you like to book a meeting or leave a message for him to get back to you?
Kevin: Could we book a meeting?
AI Assistant: Absolutely. Which day and time works best for you?
Kevin: How about tomorrow?
AI Assistant: What time tomorrow would suit you best?
Kevin: Maybe in the morning.
AI Assistant: Perfect. How about 10:00 a.m.?
Kevin: That works.
AI Assistant: Great. Could you share your email address so I can include it in the appointment?
Kevin: [email protected].
AI Assistant: Thank you. Just to confirm, is your email spelled—
Kevin: It misheard me. Don’t worry about the email.
AI Assistant: No worries. I’ll book the meeting without it. Your meeting for 10 a.m. tomorrow is all set.
Kevin: Awesome. How long is the meeting?
AI Assistant: Thank you for contacting me. Have a great day.
Kevin: That’s pretty decent. She understands, books, and even connects to your calendar. I assume it emails confirmations too?
Michael: Yes, I just got a couple of emails summarizing the message. It called you by name because I had preset your number from earlier testing.
We’re also developing a memory function, so it will remember previous conversations and customer details, making it even more humanlike.
Kevin: That’s impressive. Receptionists handle thousands of calls — it’s hard to remember every detail. Even with a CRM, they need time to look things up. The AI will have all that instantly available.
Michael: Exactly. And yes, I got the meeting request too — it came through for 30 minutes. For that demo, we interfaced with Calendly, but it can work with Outlook, Google Calendar, or other systems.
Kevin: That’s phenomenal. I know some clients use similar tools for event reminders — calling attendees to confirm, answer FAQs like parking or speakers, and so on. How much setup does this system require? Is it straightforward?
Michael: We usually customize each voice AI to the business. For individuals, I’ve simply redirected unanswered calls to the AI. For larger setups like medical centers, we integrate directly with their phone systems.
The AI handles calls first, books appointments, or transfers to staff if needed. It’s designed as an add-on, not a replacement, so it’s easy to bypass anytime. It’s simple to implement and doesn’t require changing existing systems.
Kevin: I imagine some people resist talking to AI. Have you seen pushback?
Michael: A mix of reactions. Some people don’t even realize they’re speaking to an AI — it can take up to a minute before they notice. The biggest giveaway is latency.
Voice AI converts speech to text, sends it to a large language model, gets the response, and converts it back to speech. We use 11 Labs for voice synthesis. We’re also developing the next generation with direct speech interaction. I can demo that too.
Kevin: Let’s talk latency. Humans usually have about 0.3 seconds delay, but many AIs are slower. What’s yours?
Michael: Anything under a second is good. Half a second is excellent. Our cloud models average around one second — still comfortable. Some systems are two to three seconds, which gives it away immediately. Latency tuning is crucial.
Kevin: Makes sense. I’ve seen tools claiming half a second now, so it’s improving quickly. Let’s see the next demo.
Michael: Sure. This next one is experimental — it’s a real-time AI without text-to-speech conversion.
AI Assistant (Demo 2): Hello, this is Maya from Midtown Medical Associates. What can I help you with today?
Michael: Hi Maya, I’m feeling a bit unwell and want to book an appointment.
AI Assistant: Sorry to hear that. Could you tell me briefly what’s bothering you so I can assign the right doctor?
Michael: Just a virus, I think.
AI Assistant: Got it. Do you have a preferred doctor or whoever’s available soonest?
Michael: Whoever’s available soonest.
AI Assistant: Dr. Lee has an opening tomorrow at 2:30. Does that work?
Michael: Yes, that works.
AI Assistant: Great. I’ll book that for you. Could I get your phone number?
Michael: 12345.
AI Assistant: Got it. All set, Michael. See you tomorrow at 2:30. Feel better soon.
Kevin: Very smooth. If you’re a receptionist, you’d be worried about being replaced. But this is incredibly efficient — no more waiting on hold. How many calls can it handle at once?
Michael: As many as needed. One of our setups runs 20 concurrent lines, but it could easily scale to 100. You only pay when the AI is active, so no cost during off-hours.
Kevin: That’s great. The quality’s high, setup seems quick. How long does it take?
Michael: For complex systems like medical, around three months. For simpler setups, a few days. We do thorough testing — thousands of calls — to make sure the AI says the right things and recognizes emergencies, mental health issues, or suicidal indicators. It flags urgent cases to humans immediately.
Kevin: That’s phenomenal. What kind of returns are your clients seeing?
Michael: The ROI can be huge — over 380% in less than 2.5 months. For example, one clinic with 20 receptionists misses about 10% of 62,000 monthly calls. AI handles 50% of calls, freeing half the staff and eliminating missed calls.
Setup might cost around $200K, with ongoing costs around 40 cents per minute. But the savings and additional bookings can generate close to a million dollars in net annual benefit.
Kevin: A two-and-a-half-month payback and 386% ROI is impressive. At what size does it become viable?
Michael: It depends. For individuals, around $120/month could get you an AI assistant to take messages and book appointments. For businesses with constant call volume, the savings become significant. It’s ideal for those with multiple staff handling repetitive calls.
Kevin: Amazing. You’re clearly at the forefront of this. For anyone interested, where can they find you?
Michael: Visit exiq.com.au, find me on LinkedIn (Michael LeVene), or call 0423 547 02. And if I don’t pick up — my AI will.
Kevin: Perfect. All links and details will be in the show notes. Any final thoughts?
Michael: This technology is becoming mainstream fast. Businesses that adopt early will offer better service, reduce costs, and grow faster. The longer you wait, the more your competitors get ahead. And while we use medical as an example, this works across most industries — anything except defense, which has its own challenges.
Kevin: Michael, thank you so much for your time today.
Michael: Thank you, Kevin. Appreciate the chat.
