From Zero to Your First AI Voice Agent in 18 Minutes (No Coding) transcript

AI agents are one of the biggest [music] shifts happening in technology right now. And in 2026, they're only going to get more powerful. And if you've been watching [music] from the sidelines, it might feel like you're getting left behind. Then you look at some of the examples or tutorials, and they seem way too technical. But here's [music] the truth. There is a corner of the AI agent space that's actually way simpler to build and is actually [music] driving the most real-world ROI taking place right now.

I'm talking about voice AI. [music] These are agents that can answer phone calls, have real conversations, qualify leads, and book appointments for businesses [music] without a human ever having to pick up the phone. And right now, complete newbies are selling these systems for anywhere [music] from $5,000 to $25,000. And in this video, we're going to break it all down. What an AI voice agent actually is, >> [music] >> how it works, what it can do, and finally, a step-by-step guide on building your first AI voice agent with no coding required.

All right, to start off, let's back up for a second. What even is an AI agent, and how are voice agents different to the ones you've probably heard about? Here's the simplest definition. An AI agent is a system that can think, remember, and take action. I like to think of them as a digital employee that can reason through problems and actually do things, and not like most chatbots you're probably familiar with, just generate text in return.

You've probably seen these kind of complex innate and workflows on make.com that look a little bit like this on YouTube. Those agents are working with text, processing information, triggering actions, and moving data around the company's databases. But regardless of how complex they look, every agent comes down to three core components. Firstly, you have the AI brain. That's the AI model like ChatGPT or Claude or Character AI, whatever you're using.

It's what actually does the thinking and the reasoning. Then secondly, you [music] have the memory. That's the ability for it to remember what's been said in a conversation, because without memory, every message would be like starting from scratch, [music] and it would have no clue what it's done. And thirdly, you have tools. And these are the actions that an agent can take, like [music] booking an appointment or sending an email or looking up information in a database.

And the tools are what let agents actually do things in the real world, and not just talk about them. So, you got brain, memory, and tools. That's the foundation of every AI agent you've ever seen. So, what [music] makes voice agents different? So, a voice agent has all these same three components: a brain, memory, [music] tools, but an extra layer on top or kind of around it that lets the agent communicate through speech instead of text.

So, here's how it works under the hood. When you call a voice agent, your voice gets transcribed into text, then that text gets sent to the agent where the AI brain processes it, uses tools if it needs to, and then generates a response. Then that text response gets converted back into speech, and you hear the agent talk back to you. It's essentially a wrapper around that same agent technology, but it opens up a completely different use case: the phone call.

Now, a lot of the newer models can process voice directly without the transcription step, but this transcribe-process-speak approach is how most systems worked up until today, and it's incredibly effective. So, this all sounds cool, but why does voice specifically actually matter? Well, it's because voice is actually the fastest medium of communication that we have with technology, and that's not going anywhere until we have brain implants.

So, you can think about the scale here, the amount of opportunity. There are hundreds of millions of phone calls that happen every single day. Businesses are drowning in them, which means there is a huge industry that is only just starting to take off with these AI voice agents, which anyone can get into right now. So, here's a simple example that really proves the value of AI voice agents. So, imagine we have a salon owner called Sarah, and Sarah's great at what she does.

So, she's always booked out, but she's also losing her mind over the constant calls she gets every day while she's at work, and when she's not at work. Half of them are spam, half of them are people asking questions that they could have found on her website, and some of buried in there are actual clients who are trying to book an appointment right now. So, every time she hears a phone ring, it's a guessing [music] game.

Is this actually worth stopping what I'm doing? She's in the middle of a haircut, her hands are full, and she's got no idea if that call that's coming in is a potential new client or someone cold calling her to try and sell her insurance. An AI voice agent handles all of that. It answers every call instantly, filters out all the spam, answers the basic questions from potential customers, and it's only going to interrupt Sarah when something actually needs her attention.

And this is why small businesses all over the world are paying anywhere from $5,000 to $50,000 to have these kinds of systems [music] built. Voice agents help businesses to make more money, save costs, and reduce mistakes [music] caused by human error. And you're about to learn how to build one yourself. So, now you understand what a voice agent is and how they work under the hood. But what can you actually use them for?

So, let me walk you through a few of the most common use cases that businesses are paying for right now. Starting with an AI receptionist or inbound receptionist. This is one of the most popular use cases, and it's actually what we're going to be building today in this video. So, an AI receptionist can answer incoming calls to the business, greets the caller, ask a few qualifying questions, and books an appointment directly into the calendar [music] of the business.

It can handle basic FAQs, filter out spam, and transfer calls to a human as and when needed. So, you can think of it as a 24/7 front desk that never needs to take a break, never gets sick, and never puts someone on hold. Second one is lead qualification, and this is huge [music] for businesses who are running ads. So, when someone fills out a form or clicks on an ad, the voice agent will call them back immediately, sometimes within seconds, and asks them a few questions to see if they're a good fit for what the company is trying [music] to sell.

And if they are, it can book them straight into a sales call from there. And this is so valuable because speed [music] matters for businesses here. If you call a lead within the first 5 minutes, you're dramatically more likely to be able to close them. Thirdly, we have customer support. So, instead of making customers wait on hold or navigate through a frustrating phone menu, a voice agent can handle common questions right away, and it can handle multiple calls at once, so it can scale with the business as their support needs grow.

These customer support voice agents can answer things like, "What are your hours?" or "Where's my order?" or "How do I reset my password?" And the agent can handle those questions instantly. And if the issue is a bit more complex, it has the ability to transfer it into a human with all of that already about what their problem is. Fourth and final is appointment reminders and confirmations. This is a simple one, but it actually saves businesses a ton of time and money because the agent can call customers just before their appointment to confirm that they're actually still coming.

And if they need to reschedule, it can handle that, too. So, businesses face way less no-shows, and there's no more of this back-and-forth texting, which is a huge waste of time [music] for the team. So, those are just a few examples of what's happening in the AI voice space right now. There are dozens more from outbound surveys to payment collection to internal HR systems as well. But for today, we're just going to focus on building one of the most in-demand use cases, an AI [music] receptionist that answers calls, books appointments, and can hand off to a human.

Let's get into it. All right, so let's talk about what we're going to be building here. It's an AI receptionist for our friend Sarah that does three things. Firstly, answers questions about the business. It's going to be pricing, location, hours, services. We're going to give the agent a knowledge base, and it handles those questions easily and instantly. Secondly, it's going to be able to book appointments. The agent is going to be able to check the calendar for availability, collect the caller's details, and then actually book them in on the calendar.

No back-and-forth, no missed bookings [music] either. And thirdly, the agent is going to know when to hand off to a human. So, if the caller gets frustrated or the agent can't help, it's going to transfer the call to a real person, in this [music] case, Sarah the business owner. We can build this whole thing in less than 10 minutes, and we've got a handy trick to help speed this up massively for you guys. And when you hear the AI agent's voice, you're going to be shocked by how realistic it sounds.

All right, so step one in building your first AI voice agent is heading to retail ai.com. That's the tool that we're going to be using for this video. They actually sponsored this video, so shout out to the retail guys. They have a really, really good AI voice agent platform here for beginners. As you're going to see, super quick, and the integrations to calendars are really, really, really rapid. So, I highly recommend retail.

It's actually the platform at Morning Side AI, my agency, that we use to build our client projects. So, happy to be sharing this with you guys today. So, retail AI, you hop on here, you can head to log in or create an account. In a few seconds, you'll be able to get account created, and they have a free plan as well, which is good for you guys who are just looking to dip your toes in. Now, here on the home page, you want to go up and click on create an agent.

We're going to go to voice agent. We're going to go a single prompt agent. I won't go to all these different types, but for now, a lot of the things you'll be building as a beginner would just be a single prompt agent. And with the way the voice AI space is going, more and more you're just going to be using single prompt agents. So, we're going to create that starting from blank. And then on here, just to the quick orientation, you can see a few things.

Here's where we put the prompt for our agent, which we're going to be writing in a few seconds using AI, a little trick I've got for you guys. Then we have our welcome message, so whether the AI speaks first or the user speaks first. I've got the model that we're using, which I'll go into in a bit. The voice that we're using, which we're going to be using an Eleven Labs one that sounds really good. The language and a bunch of other things here, which we'll get into in a sec.

But the other side that you need to worry about is this testing panel, which we'll be using to test our agent in a little bit. The first step we're going to do here, just because it's nice and easy, is set up our knowledge base. Now, if you remember, we said this agent was going to be able to answer questions about the business from knowledge base. can come over here and create a new knowledge base. I'm going to call this Sarah's Salon.

Then we're going to add a document. I'm going to upload a file. And in this case, I've already made the Sarah's Salon AI receptionist knowledge base. So, if I open this up here, you can see it's just got a bunch of information about the business, where they're based, their location, hours, backstory, etc. or the team, their pricing, etc. So, I just generated this with ChatGPT, but I've got it as a resource for you guys if you want to grab it.

This and something we're going to be using in a second is going to be on my free school community. You can join that through the first link in the description. You'll need to fill out a little form. Once you get that in, you can head to the classroom, and then into the resources section where you'll be able to find this video and all the resources I'm going to mention that are coming in a second. So, Sarah's Salon's knowledge base, we can add a document on here.

You'll be able to get that from the resources section on School. We're going to upload the file, save that. It's going to take a little bit to process, so we can actually hit back to our agent. You head here, this single prompt agent, and now we should be able to go to the knowledge base, add, and we see Sarah's Salon. So, we can click that now. It's going to connect that knowledge base that we just created, so that our agent has access to it and can answer questions based off the information inside it.

So, that's one of our core functionalities of this agent done. Just like that. And now we need to do kind of the trickier part, which is setting up the real-time booking over the phone. Now, Retail makes it super easy by going into this function section, clicking add, and then we can go check calendar availability through cal.com. But this is going to be the first tool that we set up, and it's the ability to check availability on a calendar.

So, to set this up, we need to hit to cal.com and create an account here. I think it's easier just to sign up with Google. I'm going to be using my Google calendar, so I highly recommend just taking one of your Gmail accounts and connecting that to cal.com because we're going to be syncing with Google calendar, and it makes it a lot easier. Cal.com wants access to my Google account. Yes, that's [music] all good. And in this case, I'm just going to go for personal use.

It's nice and free for us, which is great to play around with. I can set things up like my bio here. I'm not too bothered, if I'm honest. You guys can set that up properly. And then as you can see here, because I used a Google account, my Google calendar is already connected via cal.com. So that's going to save us a bunch of time. We can continue there. Now we're into cal.com. So as you can see, it comes with a few default meetings.

What I actually want to do first is to create a new meeting. I'm going to call this a cut and blow dry. So just a little bit of info to make this event seem real. I'm going to make this a 60-minute duration. I don't know how long that takes for a cut and blow dry, but I think maybe 60 minutes, longer for a woman, I'm not sure. But cut and blow dry, we're going to have this new event [music] type. Now we have this page here.

Once we had our event set up like this, we can head back to cal.com. You can see what it's asking for here is an API key and an event type ID. So the easiest one to grab right now is as we click into this, if you are not already, if you go back, make sure you click into the cut and blow dry. And if you look into the URL here, you will see that there's a number. Now this might be a bit tricky for some of you, but make sure you have gone to this home page here.

You've clicked on the new event you created and then go to the URL, whatever browser you're on, and grab this number in the middle. [music] You want to take that and head back to Retail and then we're going to paste that into the event ID. And then we go back to cal.com again and we're going to go to the settings here. We're going to head to the API keys. We're going to click create a new API key and we'll call this Retail.

I'm going to set this to never expire. Save that. And then copy this API key, head back to Retail, and paste it in as your API key here. I'm going to delete the time zone just for now. [music] And then we can just click save. Now you can see we have one of our functions or our tools set up and that's going to allow the agent to check the availability on our calendar based off that event that we've set up in cal.com. Now we need the ability to actually book in something on that calendar so we can come again and click on [music] to book using cal.com.

Now you should still have copied that API key from the last one. We can just paste that again directly into there. Then we can head back to cal.com again and we can just grab that event ID again. Where is it? Here. And paste that back in here. I'm going to remove the time zone and we can double check that this description for the agent is correct. So when the user asks to book an appointment, book it in on a calendar for a cut and blow dry, book it on the calendar with this tool one hour event.

All right, so now we have most of the hard work done. We can actually add in another one here which is the human handoff if anything starts to go wrong with the call and we can click on call transfer here. We can change the description to transfer the call to a human, the owner of this store, Sarah. And here for testing sake, you can put in your own phone number. If you're doing this for a real business, this would actually be the real business phone number.

We want to just set this up as a call transfer and then we can click save. Now most of the work is done. All that we need now is a prompt in this big box here that is going to stitch everything together and define how the agent behaves. That's who it is, who it's working for, how to use the knowledge base, how and when to use the tools that we've set it up with and things like its personality and really importantly, stuff like guardrails of what not to do and how to behave in certain situations that may be a little bit riskier.

Writing out these prompts manually can take forever so I use a handy-dandy tool for me when I'm creating these kinds of AI receptionists that quickly takes in all the information and generates an AI prompt for us that's going to make it function as we expect. So this is my handy little tool here and I can just fill this out with the information about the agent and then get a prompt in a few seconds. So just a quick blurb here what the agent does, where it's being deployed and why.

The conversions context here is about what it's trying to do for the business, what's the objective. Knowledge context is just about what's included in the knowledge base. And now the tools available is an important one to get right so we want to hit back to Retail and we want to actually copy these. Head back over to here, paste it in. Now we have all the exact function names that we can put into our prompt to ensure that the agent knows exactly what tool we're talking about when we're [music] describing how it should be used.

So as you can see, I've got the name of each tool and a bit of context around each of them on how they should be used such as if they're trying to manipulate the system that they should use the end call function and when transferring the call, it's going to be passed off to the owner, Sarah. Final notes here. And once you've filled all of this out and this tool will actually be available for free on my school community with all the other resources that come with this video.

So if you want to grab this, it'll be in the school community as well in the resources section. We can come down and generate the prompt. And if we scroll down, we see that we have the prompt here and we actually want to click on the formatter button here to get the raw version. Click the copy button here. Head over to Retail and then paste it into this box here and just like that, we have a perfectly formatted written prompt for our AI receptionist and that has saved us a ton of time of trying to write that all manually.

If you're following along and just want to skip all that process and just steal this prompt, I'll also include that on the school if you want to get this thing up and running even faster. So just give me through the prompt, you see it's got the personality, the environment, the tone and the goal. So the first one here we want to look at is this greeting. So the AI is actually going to speak first and we're prompting it into what it should say.

So we can go AI will speak first, leave it as a dynamic message which means the AI is going to generate it itself. Going to ask questions, check availability for the one hour cut and blow dry appointment using the exact names for the tools. Going to book the appointment, handle difficult situations by transferring it to Sarah and then using the end call to end the call politely. Got some guardrails in here and the tools as well.

So we're just about ready to give this a test. The final thing we can do is just head into the model section and pick which model we want to use cuz we are really spoiled for choice. Just because I want something fast and cheap, I'm going to go for Gemini 2.5 flash light and I'm going to select the voice here. You guys can play around with all of these, they sound really really lifelike. Let's just filter by female as this is a salon.

Accent, let's go American and let's go for someone young. Let's just pick Evie. Use this voice. One cool setting we can put on before we test this thing is actually putting a coffee shop background noise. You'll see what I mean in a second. Making sure we've saved this prompt. And now we're ready to give this thing a test. And as you can see here, we actually can't test the transfer on the web call so we're not going to be able to do that, but I can test the rest of the functionality now.

All right, now before we test it, I've just got our calendar up there on the side so that I'll be able to see when the calendar booking actually comes in. Make sure that it's worked correctly. And now we can test it. Hello. Thank you for calling Sarah's Salon. My name is Mary. How can I help you today? Are you calling to inquire about our services, pricing, or to book an appointment? I'm going to be honest. I don't like that voice.

Grace is giving the call ups. Hello. Thank you for calling Sarah's Salon. My name is Mary. How can I help you today? Are you calling to inquire about our services, pricing, or to book an appointment? >> How much does a cut and blow dry cost? A cut and blow dry with our senior stylist is $95. With our junior stylist Mia, it's $70. Would you like to book an appointment? >> [music] >> Great, yeah, I'll get one with the senior stylist, please.

When can I book in tomorrow? Like 11:00 a.m., 8:00 a.m.? I have an opening tomorrow at 9:00 p.m. Would that work for you? Yeah, that works. Great. Can I get your name and number, please? Yeah, my name is Liam Autley and my number is 021 805 9342. Thank you, Liam. I've booked your one hour cut and blow dry with our senior stylist for tomorrow at 9:00 p.m. Perfect. We look forward to seeing you tomorrow at 9:00 p.m. Thank you for calling Sarah's Salon.

Goodbye. Boom, there you go. And as you can see, there's a couple details here you'd want to tidy up and also here in the time zone, you just want to set the same time zone here as you have on cal.com as well. But overall, it's not dragging us out much further. That is how you set up an AI receptionist on Retail in just a few minutes really using cal.com and an AI prompt writer that I'm going to give you guys access to in the resources for this video.

Now you may be wondering what the next step from here would be which is to actually deploy this to a phone number. You can head back to home here and go to phone numbers, add one in. We can buy a new phone number. But in order to get a phone number here, you're going to have to sign up to a paid plan with Retail. So when you actually want to put into production, you are going to have to pay, kind of makes sense. >> [music] >> But once you have your phone number, you'll be able to go to your agent and deploy it directly to that phone line by publishing it, giving it a version name, clicking inbound, and then clicking publish.

And then anyone who calls that phone number is going to have that exact same experience >> [music] >> that we just had there. So easy, scarily easy, and scarily good as those voices are as you can tell. But that's it. That's how quickly you can create a powerful voice agent that automates hundreds of hours per month for a business. And obviously, this is just the start. There's a lot more complexity you can add like handling multiple booking types and event types, connecting to different calendars, and then testing and improving the agent over time.

But I just wanted to show you guys how accessible this is and explain the incredible opportunity you can get into right now. So if you're interested in learning more about building with AI and actually making money selling systems like this, I actually just released a full course on how to sell an AI business in 2026 even if you have no AI or coding experience. It's a complete playbook for 2026, everything you need to know to learn these skills, land your first clients, turn this into a real business from home, and it comes with a free course that includes all the templates and tutorials to get you there.

So that video is going to be linked right up here. I highly recommend you jump onto that next. It's also going to be linked in the description below. But aside from that guys, thank you so much for watching and I'll see in the next one.