This is the full transcript of Making $$$ with Voice AI agents and SaaS (my little secret), published on YouTube by Greg Isenberg. Every paragraph carries the moment it was spoken, so you can click any line to jump straight to that point in the video, search the whole thing for a word, or copy it out.
0:00Today we're going to talk about how to make money with voice AI agents. Now I think a lot of people when they hear voice AI agents they just think of these robot callers that actually are horrible experiences. Now I think that there are tons of those poorly set up voice AI agents but at the
0:19same time we're getting to a point where some of these voice agents actually sound human beings. And I think there's a bunch of businesses that you can create using voice AI agents that could generate a lot of money. So today we're going to give you a primer on how to master voice AI agents. And I'm going to give away a bunch of startup ideas, the exact workflows on how to actually create these startup ideas and frameworks. And let's just start with one just to get you excited. I want you to feel excited. And I know that when you go through this startup idea, you're going to be , wow, I could sit through this entire episode
1:01to learn exactly how to build these voice AI agents. So, this is startup idea number one. It's an AI property manager voice line. How does it work? We've all been there. We're renting an apartment. Your toilet breaks. So you end up, you call this line, 1-800-TENANT, whatever, and it transcribes in real time what you're saying. So it knows that you said toilet broken, and then it classifies the issue as urgent versus non-urgent. Based on that, it'll either create a high priority ticket using an API and then it'll dispatch a request with notes and photos. And then you actually get a text with the ETA and a ticket link. But if it's not urgent, you know,
2:01it'll just log the ticket with normal priority and then it'll text troubleshooting steps and tickets. And then you get a push call summary and a sentiment. So this whole process, you can use, you know, VAPI and 11 Labs. You can use DeepGram and Whisper. You can use a property management tool AppFolio or Buildium and just connect with those softwares. And that's, by the way, one of the things I love about voice AI agents is you can use it and connect it to outside software, repair vendor, and then you can use something Twilio for SMS. If you didn't really understand that, you know, this framework, I also thinking about it from a flow perspective. So the inbound tenant
2:48calls, you issue the category. Is it urgent? Then you create the ticket. If it's not urgent, creates the ticket and self-help SMS. And then you sync the summary to the dashboard. This is a business that someone could create and charge $500 a month, $1,000 a month, $1,500 a month. And you don't need that many customers to have a million-dollar business here. So I think it's totally possible.
3:17And I know that there's people listening to this that are going to say, well, this already exists. Or the old-school property managers are going to get into this. Well, okay, if that's the case, just go nicher. Create an AI property management voice line for Quebec or Canada or, you know, with a specific type of brand. So I think it's totally doable. Now, at the end of the episode, I'm going to go through four more ideas this, and I'm just going to give it away. People charge for this stuff. I'm just giving it away. I want you to have the sauce. I want you to build something. My
3:56goal is to build something. But I need to start by just explaining lay of the land, the foundations, a primer about what, you know, what what what is a voice AI agent, why why you should care and what are the different types of voice AI agents. So a voice AI agent is basically a software system that could have natural phone conversations with human beings that understand what people
4:21want and that take actions to help them. Think of them as digital employees with three superpowers. Listen and understand listen and understand human speech in real time. Think and reason about what you're asking for and speak and act to solve your problems. Now, there's four main types of voice AI agents. Number one, you have customer service agents handling complaints, returns, order status. Number two, you have sales and leads qualifications. That's qualifying prospects, gathering requirements. Number three, you have appointment scheduling. You can book meetings, manage calendars. And last, you have information and support, answering questions, providing business info. So, you know, that's what the four main types are. Now, if we just zoom in here, this
5:11is the way to think about it. How does this work? How do these voice agents work? Well, basically, step one is they listen to your voice. I need to book an appointment. Step two is, and this is the key point, and this is why it's finally starting to get voice AI agents are starting to get good.
5:29It's because the models are getting better. So there's an AI brain that understands the intent. So basically, when you're on the phone with one of these bots, you're on the phone with an LLM, a Gemini, a ChatGPT. So it understands what's happening. And then there's text-to-speech, step three, text-to-speech that converts the response to voice. I'd be happy with
5:53that. And step four, there's some action. So it takes an action in a business system. It books the appointment. It updates the CRM in Salesforce. It sends a confirmation. Real quick, I need to tell you about something that's helping me sleep better at night as an entrepreneur. I was running multiple companies with money scattered everywhere, zero spending visibility and making growth decisions pretty blind. I knew there had to be a better way. And my friend told me about Brax. So I switched and now I have mission control for every dollar. Spend management and visibility, virtual cards for my team, and earning cash from the same-day treasury product. I was leaving so
6:39much money on the table, but not any more. I have a financial operating system for our businesses to make decisions so I can spend smarter and move faster. And it's been a total game changer. I'm loving Brex so much, I reached out to them and asked them to sponsor the pod. If you're building something real, you need a complete financial operating system Brex. It'll help you sleep better too, and you can't be mad at that. Now what are the key technology components and what tools do we need to care about And by the way at the very end of this episode I will go through which tools
7:21the pros and cons of some of these tools and if you're technical, which tools you should use, and if you're non-technical, which tools you should use. That'll be at the end of the episode. Now, there's a few components to voice AI agents. You have the speech-to-text. So you're going to need that to convert the voice-to-text, so you can use something DeepGram, Whispers, assembly AI, an AI brain. You can use something GPT, GPT-5, I guess, Claude and Gemini. Text-to-speech, you can use 11 Labs, which I've talked about a lot on the pod, Cartegia, OpenAI, TTS. And then there's
7:59these platforms that basically handle everything for you. I've had Lindy AI on the podcast before. They do a lot of voice AI, But Vappy and SynthFlow and Retail are also the big ones. So that's kind of from a tech perspective, what you need. And a lot of people think that voice AI agents are just really just focused on, it's a limited use case. And it's just not true. You can apply voice AI agents to healthcare, appointment booking, insurance verification, home services, emergency calls, service scheduling, legal, client intake, case screening, restaurants, reservations, menu questions, real estate, lead qualification, showing schedules, e-commerce, order status, and returns in exchange. So this is, you know, we haven't even scratched the surface on what to
8:57do with voice AI agents. I want to just make one important point here. A lot of the time, actually, human beings are better than the voice AI agents. The only thing is human beings don't work 24-7 and voice AI agents can. And also, some people don't have the budget to hire 10 sales reps or 10 assistants and stuff that. So it's just cheaper to do this, right? If you're trying to save money, it's a good way to do it. I will say, though, if it's a complex request, what I to do here is, you know, I'll start a voice AI agent. But if it's complicated and the LLM could see it's
9:44complicated, you can have it transferred to a human. So you can say, let me connect you with someone who can help. And what some people do here is if it's sales related, if it's if it's close to making a sale, just transfer it to a human being. So that's just something to think about is kind of a hybrid approach. You don't need to go full on 100 percent voice. I think you can have some voice. I think agents help you and that'll get your cost down, help you scale. And then when you actually want to do the sale or you want more of a white glove experience, that's when you do
10:16transfer to human. We have a paid membership for entrepreneurs called Startup Empire. There's about 600 people in it, and it just helps people who are idea people, help them build their businesses faster. We thought it would be fun to build a happy AI agent that gets feedback from our customers because we always want to make it better and better. And we use the flow this, and it's working pretty well. and I'm going to actually call Anna, that's her name, Anna Eisenberg, AI,
10:50and call her and see what she has to say for us. And this is how good voice AI agents are. Don't take my word for it. Take Anna's. Hello, this is Anna. I'm Startup Empire's AI assistant. I'll be in charge of collecting any feedback you have about your experience in Startup Empire. All of your feedback will be completely anonymous unless you decide otherwise. Let's get started. Would you mind telling us your name? If not, just say no thank you. My name is Greg. How has being part of Startup Empire helped you or your business? It's been pretty good. I've only been a member for 30 days, and I noticed that there's some amazing people in the community, but I want
11:37to spend more time with them. I just don't know how. Could you please provide more details? Yeah, absolutely. I noticed that there was someone named Jordan in the community, and I want to hop on a Zoom with him or just get to know him better because I just think he's super smart. What did you think was the most useful or impactful part of your experience? I would say every few weeks, Tony does an AI chat where he just shows the latest stuff and what's working for him. And that has been the most interesting thing for me. If you could change one thing about our community or
12:17product, what would it be and why? I think that there's so many smart people in the community. It's just about how do I spend more time with them? Is there anything else you'd to add? No. Thanks for taking the time to give us feedback. We appreciate you. Bye. Bye. So the cool thing about
12:36it is now what we do is we actually take that feedback. we send it to an LLM. It summarizes it. It shares it on our Slack. And I can actually show you that. Let's see. Empire notifications. You can see Greg. So Greg, a 30-day Startup Empire member shared valuable feedback hiding the community's smart people in Tony's impactful AI sessions. His main suggestion is to create more opportunities for members to connect deeper and specifically mentioning wanting to interact with more community members Jordan. How cool is that? How cool is that? And now our community manager can go in and
13:14be , okay, there's this person named Greg. Let's go connect him with Jordan. Let's do more events. And everyone on our team could see. So this is a bit of the primer on what voice AI agents are. And I want to get to the fun stuff. I want to get to the ideas. I want to give these ideas, and I hope you end up building it. The second idea is an AI patient intake assistant for dentists. So,
13:43you know, you call up this number. I'd to book a cleaning. It transcribes an entity extract. So your name, your date of birth, your insurer. And then it checks the patient record. Are you an existing or new patient? If you're a new patient, it needs your insurance provider or member ID. So then it goes and creates that. And then it'll create that in a software. The popular ones in the dental world are Dentrix and Open Dental. It queries the appointment slots. It'll see if there's available times. It offers the top three times for you. You select the slot. It books the slot, plus a little buffer. It adds appointments and the notes. And then it sends the confirmation
14:27and forms links So the tech stack needed for something this is you going to use something a Vappy a Deepgram whatever system that the dentist is using Calendar you can use Google Calendar or Calendly, and then Twilio. So super simple idea, something that you could sell to dental companies all day long. It's a huge value add for people to be able to call at any hour of the day to do this. And then dental offices are good customers. They make a lot of money. There's a lot of revenue there. I could see you end up potentially selling this to one of the practice systems. Not crazy that you scale this to $1, $2 million in revenue, and they're , yeah, we'll pay you 4 to
15:186x revenue and life-changing results. So I love this idea. I think it's interesting. Someone, please build it. And then from a flow perspective, this is what it looks . The call starts and matches by the name and date of birth, the yes, no, and you get your SMS confirmation. That's idea number two. Idea number three, and I must say I stole this idea from ideabrowser.com, where if you haven't been already, go there to grab all your startup ideas. So what is it? What is it? What is number three? AI HOA hotline. So HOAs are huge, especially in the United States. There's, I think, must be millions of people living in HOAs. So let's say you call this hotline and says, when is
16:12the trash pickup? It looks through semantic search in the FAQ or database, knowledge database. It finds the answer, gives you the answer snippet. It provides the time and rules. It logs the calls and resolution and it collects the details, address and evidence. It stores the structured result or sorry, it stores the structured record and it creates the ticket and notifies the HOA manager. So this is a very simple, very, very simple flow. And I wanted to put this here because it's also, you know, it's just not that complicated. and it's about having the information in a database, you can store it in an Airtable. this is a simple, simple voice AI agent business that
17:02you can create. Use Retell or VAPI. You'll have to obviously have the knowledge database. You have to create that content, but you probably already have it anyways. And then, yeah, you just literally use an Airtable. So super, super simple. HOAs have a lot of money. is a good customer to sell into. It can scale a lot and makes the lives of people better. So that's idea number three for you. This one, this is a funny one. So I don't know about you, but I kind of remember this being a problem in elementary school. So it's called the AI school absence line. So you call this number and you say,
17:54Johnny Smith, grade four is sick today. It transcribes and extracts those entities. So it's , okay, Johnny Smith, name, grade four, got it. Name, grade, date, reason. It validates that Johnny Smith is, in fact, a student in Miami grade school. And if it's valid, it appends the absence row. It compiles and sends an 8 a.m. daily report to PowerSchool, which is software for K-12. It's a vertical software for K-12 for schools. It confirms it. And yeah, you can have it asked to
18:43repeat and clarify. And it flags the exception for the office. So, you know, could be cool, right? Could be cool. Just a very more, a very clean way of dealing with people are sick, kids get sick, and they get sick all the time. And wouldn't it be great if it just automatically updated the database into a Google Sheet? There's a daily email report. This is valuable. Now, it is harder to sell to schools. It is harder to sell to schools. So I think that if I were going to do something this, it might be one of those things where you kind of give it away. You give
19:19it away and you get them on the product. And then over time, once you've earned their trust, then you upsell them. So that's what I would do here. I would literally give this away. And then once you're in their system and you're integrated, then you can think about what are other things that you can go ahead and sell to them. So it might look a funny, simple idea that can't go anywhere, but I actually think it's a sneaky, really good idea that's a huge wedge into something that can be a lot bigger. And here's the flow for folks that want to screenshot that. Another idea, and we think we have one more idea, and then we're going to get to the different tools and the pros
20:07and cons, if you're technical or non-technical. So this one is the AI trade contracting contractor answering service. So basically an answering service. If you're a plumber, HVAC, electrician had to do a boring business idea in here. So, OK, let's break it down. How does this work? So the customer calls in and says the water heater is leaking. You do an urgency check
20:32plus address plus photos via an SMS link. And then it queries the next slot by zone and tech. You get the slot options via Google Calendar, and then you offer the earliest window. So, you know, I should have said the voice agent is something a VAPI in 11 Labs. You're checking the calendar. You're offering the earliest window. And if they accept, then you can actually take the deposit using Stripe or whatever payment processing you use. And then if it's successful, that's awesome. You can create a job ticket and attach photos. And then you can do an SMS confirmation and prep instruction. So what I love about these voice AI agents, and it's built into
21:20all these different softwares, is it makes it a lot easier for the businesses, right? So if it makes it a lot easier for the businesses, there is value that you're creating for them. And if there is value that you're creating for them, you can charge them money for it. Now, how much could you charge? I don't know, a few hundred dollars a month, maybe $800, $500. Tough to say exactly but I do think that they be willing to pay for it and on a monthly basis So I this idea I also think that it interesting that you know obviously that you can connect payments right people,
21:58that's the cool thing about voice AI agents and the integrations to these platforms is you can literally collect payments 24-7. It's kind of the internet, right? it blew people's minds that you, you know, when the internet came out that you can go on the internet and people can put in their credit card and they can buy a digital product or physical product and you can wake up every single morning and you're getting paid. Now, the same idea could be applied to all these voice AI agents. You could literally just wake up and just, if you own an HVAC company or you're a plumber, oh my God, I have these four or five new customers that came in, which is super,
22:40super awesome. And here's what the flow looks over here. Again, I think it's really interesting that you can upsell people using these voice AI agents, offer an annual maintenance plan, and then tag them in the CRM. Now, the last idea is a bit grim, but, you know, it's basically this idea that, you know, an AI funeral home intake line that's compassionate and compliant, of course. So a family caller calls the line, reports the passing of a loved one and requests assistance. I just put this in because I think it's important to note that you'd want to have the VAPI or whatever voice AI agent you use that's super soft tone low pace empathetic prompts no rush I'm here to help
23:37things that putting it in there and of course if people do want to speak to a human being you should have that option for people especially during that time So it transcribes the entity and intent extract. So, you know, the name, the time, location, maybe, you know, their religion, any preferences they might have. It creates a preliminary case file. They gather key details, pick up location, contact preferences and notifies the on-call director with a summary, confirms callback window. It provides an ETA and what to expect, document, checklist, PDFs, anything that, and then sends any compassionate resources, confirmation, and maybe a phone number of or many phone numbers of the people they can speak to if they need any support, if they want
24:26to talk about anything. I have here on-call director human. It's important to have that human in the loop. This is a hybrid idea that just is an intake form for a space that a lot of people aren't creating software in the funeral home industry. But if you can make the lives easier for that business and for the people running it so that they can focus on their job and stuff that, I think that could be great. So, you know, that's the last idea that I want to give. And I think a lot of people, hopefully a lot of people went through some of these ideas and are , wow,
25:12this is, I could do something this. I could see myself building a voice AI business, either hybrid or fully voice AI. and I just don't know which voice AI platform to start. So, you know, I can't answer that for everyone. But what I will say is, you know, do you code or not? Well, if you don't code then and your priority is speed and quality, then I would say, you know, something SynthFlow, which is about a 30-minute setup. It's a visual builder. It's about $29 to $450 a month, perfect for an MVP. If you want quality, something retail, retail AI. So it's natural conversations, handles interruptions. It's usage-based. It's a bit more of a premium experience. Now, if you're a
26:06bit more technical, and you don't have to be super technical, but a bit more technical, I would say using something Vapi. I've used Vapi for a few use cases, and it's been really great. OpenAI, TTS, Whisper, GPT, Mini. But, you know, that's in the less to 50 cents per call, I believe. So it does add up, right? More than 50%, there are some more expensive tools that you could be using. But yeah, I think for the most part, you know, using something Vappy or SynthFlow or Retail is kind of where you want to start. Vappy is a little more technical, but if you can figure it out, I highly
26:55recommend Vappy. I'm not, I don't work with them. And I'm not sponsored by any of these tools. But I do think that having, you know, I don't know. I find that if you can learn how to use it, it's worth it. And it's just you have to be a little bit technical. You have to be a little bit technical. So I should also mention that I'll include in the show notes how to actually get access to this Excalibur. I know that's a big question that people have. They see me do these and they're , hey, I need this. I need this. I don't want to just screenshot it. So please
27:34and comment before you grab it. Just do a quick and comment and then grab the link from the show notes. And I can't wait to see what you build. There you have it. This is how to master voice AI agents with some startup ideas in there. Hopefully now you have a good foundation of what is happening in the world of voice AI agents and how you could how you can partake in it. That's the magic of voice AI. And I know it's not perfect, but voice AI and connecting it with software and the automations that you can do. It's really, really something special. If you can see past, you
28:14know, some of it is janky, of course. It's only going to get better. And if you're here, you're early. So this is just scratching the surface, of course. This is the tip of the iceberg. If people want me to go deeper into any of these topics, maybe doing a tutorial on something Bappy or one of these platforms, just let me know. I read every single YouTube comment. I see every tweet that is tweeted at me. I love when people the video. It gives me a sense to know if I should create more stuff this. So please, please let me know. I live to serve and I want to see you build something
28:54big and great. So thank you. I'll see you in the comment section and have a creative day, friends.
No line in this video contains that word.
Where these words come from. This is the caption track YouTube holds for this video, uploaded by the creator. We read it, tidied the line breaks and laid it out so it can be read. The plain text version is at https://viewrankai.com/tools/youtube-transcript/pYsK9dfUnYA.txt.
All rights in this video belong to Greg Isenberg. Watch it on YouTube. If this is your video and you would rather this page did not exist, tell us and we will remove it.