Million Dollar Days

The Rise of AI Agents

• Robby Choucair and George Passas • Season 1 • Episode 84

Share episode

Send us a text

Step beyond the ChatGPT basics and into the revolutionary world of AI agents – the digital workforce that's about to transform how businesses operate forever. While most people are just getting comfortable with simple AI interactions, forward-thinking entrepreneurs are already deploying sophisticated AI agents that function like full-time employees.

This episode peels back the curtain on what's possible with AI agent technology right now. Discover how these digital workers can access your email, calendar, CRM, and other tools to autonomously handle tasks that previously required human attention. From scheduling meetings and responding to emails in your voice to generating proposals and making outbound calls, these agents represent a quantum leap in automation capabilities.

The economic implications are staggering. Tasks that might cost a business $100 and several hours when performed by humans can be completed for pennies in minutes by AI agents. They never need sick leave, don't complain to HR, and work around the clock. As we discuss in the episode, Sam Altman, CEO of OpenAI, predicts there will be more AI agents than humans on the planet within just 24 months.

We break down the practical aspects of implementing this technology – from choosing the right platforms to training your agents with the proper knowledge and connecting them to your existing business tools. We explore real-world applications including how AI agents can prepare sales calls, lead generation, quote preparation, and even communicate with each other to complete complex workflows without human intervention.

For business owners wondering how to stay competitive in the coming years, this conversation offers a glimpse into what will soon become standard operating procedure. The future won't be about whether to adopt AI agents, but how quickly you can implement them before your competition does. Ready to meet the newest members of your team? They're digital, they're tireless, and they're about to change everything.


George:

One day we're not going to do this podcast and it's going to be just the electronic AI version of me.

Robby:

I can't wait.

George:

Think about it all the time we'll see that's going to be sad. Nah, we'll still do this. People will get upset that it's not AI and that we're actually talking.

Robby:

I'm like humans. Fuck, I came here.

George:

Yeah, what the fuck, I came here for the AI. I didn't come here to hear two podcast bros talk, as I've been called before, podcast bros. Yeah, we've been called podcast bros.

Robby:

Don't you read the comments?

George:

I read them multiple times. We read all the comments, every single one of them. All you haters out there, that's why we keep coming back. We probably wouldn't do the podcast if it wasn't for you.

Robby:

Yeah.

George:

We live to piss you off that's what we're here to do, um ai so a couple weeks ago, we're having a chat and you're like hey man, have you heard of this? I'm like what the fuck is that like in the movies? Yes, I've heard of that, but not in real life. And I want to touch base with you and have a discussion with you, because you're a lot more educated than what I am and I guarantee, if you are listening to this right now, 90% of you are not educated in this space.

Robby:

Oh, dude, I think it's more than 90.

George:

Maybe 98.

Robby:

I'll give it 2%. I would say I'm not yet an expert in this.

George:

Not an expert, but you're educated. You're more educated than most.

Robby:

You understand the concept and what it can do. Yeah, I've built an agent.

George:

Oh, we just dropped it. I haven't said what we're talking about.

Robby:

Yet Don't care, you ruined it. I can't help it. I'm going to turn the podcast off. I'm done. We're in the time where speed matters.

George:

Yeah, agents, okay, not just AI. And whenever anyone thinks about AI, like what's the first thing you think of, what's the first thing they think of Chat, gpt yeah dude, the only thing.

Robby:

Okay, so just before I walked downstairs and one of your employees says hey, man, chat GPT just saved me three hours. Blah, blah, blah. He's like this AI stuff's sick. And I just looked at him and I'm like that is the most basic of AI. Like I said you're going to freak. And then I walked halfway up the stairs and I said freak when George fires you because he doesn't need you.

George:

Heard it here first, but it's the equivalent of someone coming up to you and go hey, I just did this thing where I typed into a box and it's called Google and then it showed me a whole bunch of stuff.

Robby:

But that would be cool back then.

George:

Think about it. That's what I mean. Think about it.

Robby:

Think about it. Imagine you're like how do you find something out? Let's drive to the library and find the encyclopedia and we've got to get the right letter because they're all in separate books, because there's so much information and it's not here. So we'll come back next week and then, hopefully by next week, we can find out what that thing was.

George:

Or you're like hey, we need to paint this wall. Okay, Let me just go to the yellow pages see if I can find a painter.

Robby:

Yeah, An AAAmasterpainter comes up and then you have to go find the book, sit down with the book, flick through the book, find the person, contact them, whatever, and it's like you can't believe it. This is all on one thing now you just go here.

George:

It's this Google thing. You go type in your seat. What a funny word.

Robby:

Yeah, and everyone's like whoa no one had heard the term Google Blew my mind Do you know what Google is?

George:

Zero, yes, zero idea.

Robby:

You're about to blow my mind. Oh man, I had this fact in my mind and I'm like I'm going to drop this on the pot. It's going to be sick. Is it about Google? No, but I'll tell you what a Google is. A Google is a number, oh cool.

George:

Yeah, it's a Google.

Robby:

I'm going to Google what Google means.

George:

I'll show you right now.

Robby:

You think this is a game.

George:

What's a Google? Hey Siri, what's a Google? It doesn't work.

Robby:

It asked me if I wanted to use chat GPT. Yeah, that's because, there you go, anyway, if you can. I don't know if you can zoom into that. If you can't sort your shit out, yeah. But there you look.

George:

I believe you, I don't doubt, there you go 10.

Robby:

Not a joke. 100 zeros. That was the first time I ever heard the term Google.

George:

It was in oh wow, yeah, no.

Robby:

First time I heard no.

George:

I can't remember that far back.

Robby:

I'm 40. Yeah, I don't know. I found it fascinating, I'm fascinated. So yeah, that was that, it's the equivalent. And now this is this.

George:

You know people are going and we did this at events two years ago when we did our first event, stood up in front of a crowd and you're like, tell me, put your hand up if you've heard of ChatGPT. And you had like three people put their hand up and you do the same thing today. So put your hand up if you've heard of ChatGPT.

Robby:

It's like everyone put their hands up. I made the first ChatGPT video like six days after it came out.

George:

Was it that early?

Robby:

Yeah, I got it early, people had never heard of it. Yeah, I think I heard of it yeah.

George:

And then you had done it and then it's like you and then you told me I have to check it out. Yeah, yeah, and then I set up an account.

Robby:

Sorry, maybe it wasn't six days. Maybe it was six weeks, yeah, but it was six something and it wasn't six months. Second, but nonetheless, now you go and you type into this thing and it talks back. So if you look at it from a technological, like, let's just lay it out. Are you a visual guy?

George:

I'd say so. Yes, yes, I would be, because if I'm trying to recall stuff, I look like up and right, Isn't that visual? You're looking in the right part of your brain when you're thinking about it. You're doing it right now. Yeah, but I think everyone does that. No, I think some people might look left, and I think that depends whether you're no left and right, one's creation, oh well, this is the theory.

Robby:

One's Creative, creative, logical, one's no One's creation. Like you're creating the thought and the other one's remembering. So one's looking back and the other one's usually, if they look left, they're lying. Yeah right, oh, that's the theory. I don't know how well. I've tested it. But back to what we're saying. Let's lay it out on a board. Okay, so you, as a human, now you go and talk to chat. Do you call your chat anything Chat? Do you call it chat?

George:

Me too, hey chat.

Robby:

Yeah, me too. Hey, chat, Do you message it good morning.

George:

I haven't done that, no.

Robby:

Yeah me either.

George:

Good morning.

Robby:

I say thank you, of course, and please, and your majesty and everything, just don't kill me. So right now there's a human. Human talks to chat. Chat responds. That's the whole thing how it works, right? I talk to it, it talks back.

Robby:

So imagine talking to a human that can only talk and you're like, hey, blah, blah, blah. And then they say da, da, da. And you're like, okay, cool, like this human talks, that's great. Okay, cool, like this human talks, that's great. And then imagine, all of a sudden now you're like, hey, here's a computer, it's got all these tools on it. You've got my CRM, you've got my email, you've got all this. And now you know how to think. So you know how to reason with what I'm telling you and take actions based on that, because you can do that. But you can only talk back before.

Robby:

So if you go to chat GPT now and you say, can you check what I've got on tomorrow, you can't do that because it'll turn around and say I don't have access to your calendar, right? So the next step is you give it access to tools and then you can say, hey, what do I have on tomorrow? And it says tomorrow you've got this at 3pm, this at 2pm, this at one o'clock, blah, blah, blah. And then you can say, cool, what's the weather tomorrow for my appointments? And they'll be like tomorrow, the weather's looking like it's going to be this and that, because you've connected it to your email, your calendar and your weather app. And then you can say, cool, can you just send everyone I've got a meeting with tomorrow? Can you get the details from the CRM and just email them and just confirm that they're going to be able to arrive if they haven't accepted the invite? So it'll go check who hasn't accepted the invite for your things tomorrow and then go send them, grab their details from the CRM and then email each individual one by one.

George:

A personalized message coming from you Right, this is now.

Robby:

so this is basically at the point where you can say, okay, cool, if I've got. What does a human at a computer do? Say what. What does a human at a computer do? Say what. What does a human at a computer do? They sit down, they're connected to the internet. Yes, get information and they are able to use a bunch of tools to get stuff done. Yeah, anything a human does sitting at a computer can now be replicated by an AI agent. By an AI agent.

George:

So that's how you've got to think of it, then, don't you? Then think about how many of you sit at a computer and do stuff?

Robby:

Yeah, and then if there's a step in the process where let's just say, for example, part of your role requires you to do something physically, that will probably be the only part that you need to continue to do, and every other part will be replaced. Yeah, part that you need to continue to do, and every other part will be replaced yeah, because it'll be in an automated process that can be done by agents or a team of agents so you might have.

George:

So can you have one agent that rules them all? Yeah going to. Yes, reference for lord of the rings, because I know you're a massive Lord of the Rings fan. Huge One ring to rule them all. Yeah, because you're going to have many agents working for you, so you'll have one that does your emails.

Robby:

So can I ask you a question? Can you hire someone to look after your team? Yeah, why can't you do this?

George:

Okay, so then, who would you be communicating with? Would you be communicating with the specific agent, or would you communicate with the one that rules them all?

Robby:

Okay, so let me ask you a question If you hire someone to manage your team, do you communicate with the person that manages your team or do you communicate with them all? Depends what mood I'm in, but yeah, so you can do it depending on what mood you're in?

George:

Yeah, so I could be, but from an agent point of view, I would want it to be one thing that I speak to, not multiple things. So you name your agents to make it easy, I guess, to track what they do. Shout out to Reggie Reggie, let's crack in, and Julia for me, and because Robbie's building agents, for us as well. But so we got. Yeah, I just want to be speaking with Reggie, you know what I mean. I wouldn't want to speak to Reggie, julia, frank Vincenzo.

Robby:

Yeah, that's up to you. You can do that in real life too. You can do that. You've got a GM in your business. You could just talk to your GM.

George:

You could just talk to your GM. Again, you're a massive movie buff, but I think about Jarvis from, say, iron man, where he walks into the room and he'd be like Jarvis, can you calibrate this or do this? Or blah, blah, blah, and he just talks to it like he would talk to a human and similar to what you said hey, what have I got on tomorrow? I said you've got meetings here. Cool, can you push that back? Because, like Dev and that's what I used to say like ai was never, wasn't at that level when it came out, like they didn't exist. Well, that's how I saw it and I think that's where ai is going to be really powerful, when you can start talking to it to do those things for you and you walk in. It's like you walk in the morning and you're like hey, hey, frankie, like what do I got on today?

George:

it's like, yeah, like today, perfect example, I had a Zoom meeting scheduled for 7 for 12 12pm. I actually called my PA and I said, hey, can you tell them I'm going to be 10 minutes late? And she's like, yep, cool, no worries. So she emailed him, said, hey, 10 minutes late, we're going to start at this time. She's like, yep, no worries, but you know, do that for me.

Robby:

You would.

George:

Say hey, reggie, can you I've got my 12 o'clock meeting. Can you let them know I'm going to be 10 minutes late. So, yep cool, no worries, just send them an email and it does it. It will do it. That's now Someone. So okay, that's powerful shit Right now.

Robby:

So okay, I could go now talking to your device and be like hey, email, robbie. So this will be connected to everything you do. So it's connected to your CRM, which has all your contacts.

George:

It's connected to your email.

Robby:

Would my CRM have just every single email that I get?

George:

Yeah, we can integrate that. You can do that, yeah, so I don't save my emails, they're just kind of there they feed in.

Robby:

Yeah, so it's all integrated. So as soon as someone emails you, they're in the system. Yeah, cool, yeah, so you can. You can go and say, hey, can you email uh robbie and just let him know that, uh, I want to do a podcast tomorrow and send him a calendar for 2 pm, and then it will go and formulate an email based on how you've trained it to talk. Yeah, and say, hey, robbie, blah, blah, blah. George asked me to email you. He wants to do a podcast at 2 pm. I've already sent him by at Libby no 4 Good.

George:

Yeah, and then does it do it instantly.

Robby:

Yes. So when I say okay.

George:

So when I say like what's the? It doesn't have to grab the mouse on the new email, then click on here, then write the email, then type it out. So could it do it done? Do you know? Maybe not done? Can you do how you do this, do that, can you do this, like will it take five minutes, will it take 30 seconds, like how it depends it depends on how they're trained.

Robby:

Yeah, so if they have to assess there's a couple of things, if they're well trained and they know that, okay cool. I've done this task before and I know I need this tool, this tool, this tool. It's like, let me let me put it this way I got someone who's not to send an email quicker. This is the exact same thing. You know what I mean. And people are asking questions that you could ask the same question about a human and say, like, well, who's going to be quicker here? But if you're asking the question of a well-trained one versus your PA, the well-trained one wins every time. Every time, it's not even close.

George:

Yes, and this is what's cool too, can.

Robby:

So then and this is what's cool too Can you type as fast as ChatGPT.

George:

No, not at all It'll come up with the answer. Then it'll be, and this is where you're also. It's also pretty cool because I'll get Julia, my AI agent, and say, julia, can you message Robbie and find out, you know, if he's available tomorrow to do a podcast? And she's like, yep, no worries. And then Julia actually messages Reggie, which is your AI agent, and they speak together. They're like, hey, reggie, can you check if Robbie's free tomorrow because I need to see George wants to do a podcast at 2 pm. And then Reggie's like, cool, let me check his calendar. Or goes through, checks your calendar and goes actually yeah, robbie's available, I'll just confirm with him and get back to you and then just goes back.

Robby:

She goes cool, no worries, let me know and I'll send you a meeting request. And she goes. You get a message on your phone or a does a talk and says hey, robbie, are you good tomorrow at 2 pm? This is what people are getting wrong. It would be like a human.

George:

So how would a human contact if, if it will call me, huh, it will call me or message me or something.

Robby:

So this could do the exact same thing so it could call you it could whatsapp you like hey, are you good? Are you you good tomorrow for 2 pm? It's available in your calendar. Just wanted to confirm.

George:

Yeah.

Robby:

And you'd be like yes, and it's like, all right, done. Boom Goes confirms it was the other person, it's done. Yeah, you know what I mean.

George:

And then you'd be like, okay, I want that cool. Yeah, I say, okay, thanks, robbie, that sounds great. Can you book it, make me and give me an agent? Like, how do I get an agent from?

Robby:

you, yeah. So the first thing you need to ask well, send me a dm, but the first thing you need to ask is what do I want to do? Because there are. So I think let me, let me, yeah, because this is I've asked. We went to a handful of clients four to be exact and said, hey, this is what we're doing. This is new. It's recently starting to blow up. We're going to be super early to market. You'll get in at a fraction of the price, but it's going to be a beta model. You know what I mean. You're one of the people we approached. We approached several other clients. If you're listening to this, they know who they are and the biggest friction point we've had is people don't know what they want.

George:

Yeah.

Robby:

They're unsure, because you can't think about what you don't know. Do you know what I mean?

George:

And that's going to lead into my next question. So it's like, okay, I start using it to do my emails. Yeah, all right, let's just start there from an administration point of view, because I actually get 100 emails. Oh, come on, jake, I get 100 emails a day if not more.

Robby:

So can you break down your filtering process for your emails? Yeah, I reckon I could.

George:

Yeah, I reckon I could.

Robby:

Yeah. So if you can break down your filtering process and we can train someone to filter it in the same way, they could get filtered and responded to. So not just you can set rules in Outlook to filter your emails, you can get someone to respond. So you can be like, hey, this type of email I want you to respond. You know what I mean. If the person's not in our system we've never heard from them before, blah, blah, blah.

Robby:

Or they're just, you know, reaching out because they want to tender a job, respond to them, just say we'll get in touch, save the details into the system, let's record them as, whatever their trade is. If they haven't mentioned their trade, google it or ask them and it'll do all that for you. And then you could say these emails I want you to provide. Say if, provide. I say if someone emails us about the job, provide them an update as to where the job's at. We'll connect you to our project management software so you can see where the jobs are up to, where they're up to date, and give them the updates. I don't want to talk to them. You do that. And then if my mother or father or anyone in my immediate family email, forward it through to me, because I'll respond to them yeah.

Robby:

I mean wife and kids, blah, blah, blah. I'll take that, but everyone else you sort out and it'll do that, yeah, and it can talk to them the same way. You talk to chat gpt, right, but you train it the same way you train, and you don't have to do chat gpt or open ai or anything like that. You can do any of them. You can go do google one, yeah, you can do claude, you can do gemini, you can do deep seek, whatever it's called. Um, there's heaps, but they are all. They're all. Just, they're referred to as llms, so large language models.

Robby:

It's like a brain, right. So from the ai agent space, that's the brain. Then you teach it what you need to teach, which is a knowledge base, and he gives a bunch of tools, just like you do with the employees. You get them in they've already got a brain, hopefully, yeah. And then you get them in. You, they've already got a brain, hopefully, yeah. And then you get them in and you're like, hey, this is how we do this here. This is what we're about. This is what PASCAL is, this is our brand guidelines, this is our vision mission. Blah, blah, blah. This is where the printer is. You know what I mean. This is where you go to toilet.

Robby:

AI agent need annual leave and sick leave and public holidays and shit like that. Yeah, they'll take you to fair work everything they need overtime, dude no annual leave, no, sick leave.

George:

Okay, so I get Julia to come in and she's helping me with all my shit.

Robby:

And.

George:

I call her and I say hey, do this, do this, so on and so forth. And then, as I'm working, I'm like oh, now I need it to do this. So, whether it's Julia or Gregory coming in, let's call it Greg, not Gregory, let's call it Greg. Gregory sounds like an old man. Yeah, greg sounds a little bit younger. Okay, greg. So I've got Julia and I say okay, cool, now I need someone to arrange to sort out all my events. Can't Julia just do it all, instead of getting a separate AI agent?

Robby:

Yeah, yes and no. So they do have. Can one of your CAs do it all?

George:

I mean, yeah, if they were really good. But we're talking about an artificial intelligence. This can take on a lot. Maybe an individual person can't take on that much.

Robby:

Yeah, but they also have some level of capacity. So I think they can only run so many multiple things at any given time. They can actually multitask, actually.

George:

Yeah, exactly.

Robby:

But I think there is only so much that they can do at any given time. Now, the whole premise you want, though, is you want to build it so that they are constantly doing, constantly filled with work, so just like you would it in business. When you first the first person you hired, they were probably wearing multiple hats.

George:

Yeah.

Robby:

Yeah, you probably like a bit of this, a bit of that, a bit of that, a bit of that of that. Now, when you've got a thousand employees, you hire more specifically. You're like, hey, your job is video editing and that's it. I don't want to see you doing anything else apart from video editing and on your break, you know. I mean, I don't want to see you editing photos. I don't want to see you scheduling posts on social media. I don't want to see you, uh, trying to build an ad campaign. Your job is video editing. You're either editing or getting better at editing, that's it.

Robby:

This is the same sort of concept, because you can get as many as you want. You can go from what they're talking about at the moment there's still very limited information on this, by the way, like online, but what they're talking about at the moment is they reckon in 24 months, there is going to be more AI agents than humans on the planet. More AI, think about that. That's seven, eight billion more AI agents than humans, because you will go from having 10 staff to having four staff and 78 AI agents, because now, all of a sudden, your team has had 80% of its job replaced. It's like all the manual, repetitive stuff that they do over and, over and over again. They don't need to do anymore. There's now a tool AI is a tool. Dude, there is now a tool that can do that for them. Like the same way, and having an employee is a tool. This is an employee that lives in your computer.

George:

Yeah.

Robby:

Or lives in the internet world yeah, as I like to describe it. But most people can't even fathom how they would use it, because they still can't comprehend what they can replace, because we're so used to doing things in a particular way.

George:

That's right. Yeah, I mean, even for me, as you said, I keep reverting back to my primal instincts of sending an email by myself or doing that task by myself, and you're like, no, no, the agent can do that for you.

Robby:

Yeah, the agent can do that for you. That's the next step. That's sorry, that's now. Yeah, so it should be like hey, email George.

George:

Okay. So back to what I was saying. Okay, cool, the business is coming and says all right, I want to start embracing this AI shit like this AI agent stuff Now. I want to start embracing this AI shit like this AI agent stuff now I want to embrace it. I get Julia. I want you to make Julia for me to do my memos.

Robby:

Julia with a G.

George:

Yeah, it has to be with a G, clearly. Yeah, g-i-u-l-l-i-a. I could spell it however I want.

Robby:

Yeah.

George:

She won't even be upset.

Robby:

I don't know, no, not yet.

George:

Not yet. Give it time. Don't piss them off yet. Oh, it's probably okay to piss them off now, but they'll just remember in the future. Yeah, they can't do anything about it now. They'll remember me and you're like when they send an email to your client telling them to fuck off. Hey, this is. This is Julia. George just wanted me to tell you. I overheard him talking about you today, Dude that.

Robby:

If you don't believe me, here's the recording. Here's the recording and here's the video. And then I even said are you talking about this person?

George:

And then he said yeah, fuck that guy, Fuck that guy, this is their flyer, and then she blackmails you that I'm about to hit send.

Robby:

Yeah.

George:

About to hit send. I need a boyfriend. Make me an AI boyfriend. Yeah, all right. All right, relax Sideways. I don't even know what I was going to say. Okay, yeah, so I'm a business and I want to start embracing these AI agents. Okay, how does it work? As far, do I have to go to a website and log on and download something? Is it magically uploaded into my computer like a virus? Like how does it? How do I get the agent? Like where is it an app? Like what is it? Where do I go? Okay, there's Julia. And then do I, do all the other agents live on that app? Like how do I access all of these things? So when I'm talking to my phone, I'm like, hey, julia, and she's like, yeah, hey, how are you, george?

Robby:

Good, x, y, z, yeah, so there are platforms AI agent building platforms, the same way there are data platforms. So what's the best part of here is like we're about to completely knock CRMs out the park, and then some people still don't know what a crm is yeah, it's so funny. Yeah, but what's a crm?

George:

yeah, client relationship manager. But what is it? It's a. It's a point of entry for all your information and then being able to store it and communicate. But what is it? It's a management tool yeah, no, no.

Robby:

But if you were to break it down and take away what it does, what is it? It's a software, it's a piece of code that works in a particular way, like a platform that you go on that stores data. Yeah, that's all. It is right. It's a platform you log on and it just has all your information. Yep, yep, yep, essentially, yep, this is the same thing. Okay, so then I would get so would it come through the?

George:

the platform would be the website, so I'd log onto a website.

Robby:

Yeah, yeah, you'd give me user details for a website.

George:

Yes, it all lives on the internet. Okay, so then I get that when I get into work in the morning, yeah, okay, do I just turn on my computer and say hey, julia, how are you?

Robby:

We, however you want, I can make it so that you can talk into your phone.

George:

Yeah, instead of saying instead of addressing Siri or chat, we can connect it through WhatsApp so we can give it a dedicated number.

Robby:

Yeah, connect it through WhatsApp. We can give it a landline number, a mobile number. Yeah, we can set it up so that you talk into a telegram. You know what I mean. So you can be like hey. So you can be like hey, and you've got a dedicated thing. And you know everything I send through here is going to this, to my human. Do you know what I mean? So you can talk to it directly. But, yes, it lives on the internet, on a platform. Platforms that are out now are like Relevance, relevance, ai, bland, ai, voiceflow, vappy, there's heaps, yeah, yeah.

George:

NADN N8N. N8n is a really popular one. This is where you build the agent.

Robby:

This is where you build the agent and it's all stored.

George:

So now you've got an agent that does let's just say, you've built an agent for that particular purpose just managing boxes. Yeah, that's it. That's all Julia's going to do. Yeah. You going to do yeah, you're going to send it to me.

Robby:

I'm going to get it.

George:

Yeah, Now do I need to? Is there a backend thing that you have to? Is there prep that you have to do? Do you have to say, okay, you're using Outlook, I've got to connect you to George's Outlook. Yeah, Do you do that? How do I do that?

Robby:

Do you get?

George:

instructions no, so we build them for management.

Robby:

So you need to make sure that things are up to date, because if it loses connection with the tool, it can't turn around and say one sec. Let me pick it up.

George:

If Outlook is down, you can't send emails, even with your employees. If the internet goes out, they can't use the tools to send an email. So will the agent be able to self-teach, or do you need to do that on your end? So say, if I go, okay, julia, I want you to start his outlook, just organize any emails you get from these people. Just delete them. It's spam. I just want you to delete them straight away. Any email that I get from my team say, hey, you can answer it based off what I would say. This is the rule. So on and so forth. Do I need to tell you that or can I? Similar to how, when you talk to ChatGPT, it gives you the answers by the prompts that you give it? Do I have to train the AI agent or is it you that does that? Like I'll call you up and say, hey, I want it now to do this. I need it to also connect to my calendar and the weather, because I want to do 80% of my meetings outdoors now.

Robby:

So certain aspects it will self-learn, and then there's particular aspects where it would have to be done on the back end, like it's not going to be able to go and connect itself to an app. You have to get someone to go and connect it to it.

George:

So then could I have Julia connected to my emails, but also connected to WhatsApp at the same time?

Robby:

Yeah, and then you could WhatsApp Julia.

George:

Go into the PASCON. You know WhatsApp group the operations group and it'll be there. Yeah, so I would add her as a person.

Robby:

Yeah, yeah, so it would sit there as an individual.

George:

Yeah, and then I could say hey, can you. Just there could be 100 and 200 messages that go there. Yeah.

Robby:

And then you could message her separately and say can you give me a summary from that group chat?

George:

Oh my God, marry me.

Robby:

Do you?

George:

get what I'm saying.

Robby:

Should I tell my wife about Julia? This is what I'm saying. Of course I should. Yeah, I don't have to do the thing anymore. Yeah, it'll do the thing. Do you know what I mean?

George:

The same way it does, because this is what I mean right, you could build this agent now and whilst it's tailored for me, it can be tailored for it could be used by everyone, as in all these other individuals and companies.

Robby:

Now, yeah, the template, yeah, but the training no the training.

George:

I have to do.

Robby:

No, it's got to be fit to PASCOLON.

George:

Yes, of course, of course, but I'm saying specifically for the time being like the emails, the WhatsApp, like this agent, you've built it now, yeah, yeah, I'm just thinking of this fucking bubble on the internet or on the platform. Yeah, you can use that bubble and plug it into another company.

Robby:

No, why? So you can use the template. So, I want you to think of it like this If there is okay, you built this house.

George:

So you can now make me the same house.

Robby:

like that right, yes, no, you can't.

George:

I could Do. You know what I'm saying.

Robby:

Yeah, yeah, I can build you that same house, I can probably build it much quicker. Yeah, I've kind of got the frameworks and I know the stuff and we've got all the design stuff and I've got leftover supplies.

George:

I've got that house. I want to build that. If you want that same house, yes, I can build it for you. So I have to go out and actually physically build the whole thing from scratch. So you have to build a brand new agent from scratch.

Robby:

So, no, no. So the same way you can template, so the same way you're not building the house from scratch because the design process is already done.

George:

Yeah, so the design process is done. Yeah, yeah.

Robby:

We can take the design, we can upload, so that the platforms allow you to upload templates. So I can say, okay, cool man, this bot that we've created that does all these things, turn this into a template, and now I can clone that because I.

Robby:

You still need to make it. You, yeah, does that make sense? So like this now I've trained this for George. What if so-and-so down the road doesn't want the same thing as George? Yeah, he's like no, no, my emails, I want this, this and this. And it's like, okay, we need to make it. Do that. Okay, the same way. Someone come in and be like what color paint.

George:

So the plans are the same. The plans are the same, but then the execution of building it has to be-. The body is the same, it still has to be built. The frame is the same. So does that then say Does that then mean it's going to take you a long-. So is it hard for you to scale? No, what's why Everything is scalable, I know, yeah, that's what I mean.

Robby:

So could you have an agent that?

George:

builds agents. Yeah, I don't think they have that yet. Yeah, but you know what I mean. So I'm trying to say, like how could you sell? Sell 10 000 of these agents to people. I don't know you don't know that answer yet.

Robby:

Yeah, that's why I'm just thinking there's, there's, there's.

George:

It would be a long process if you have to build 10 000 agents now, yeah, that's a big process. Oh, if you, if you have 1,000 employees and you want all your employees to have a Julia, can that work too? Why not no, I'm asking you. I don't know.

Robby:

Yeah. So if you've said hey guys we've now rolled out a new AI agent.

George:

They'll be numbered. This is our agent. Her name's Julia 37. So each person will have their own Julia number.

Robby:

Yeah.

George:

How many names are you going to have? I'm looking at it from what?

George:

the agent does, yeah, so by putting them in a box. So Julia is our email and organising agent. It's like your own personal assistant. So, everyone, this is Julia, julia, this is everyone. Julia, this is everyone. Yeah, everyone gets to meet her and has those. So could she be intertwined with all the other Julias, like Julia 14 and Julia 37? Yes, and know what these people do and how they talk and all that sort of shit? Yeah, and then learn from that as well, like, would she learn about the company, about how everyone writes emails? That as well. Would she learn about the company, about how everyone writes emails? Then maybe I could have Julia number one, because it's the number one. I say Julia, who's fucking up? Send me a report. Who's not doing this? Who's?

Robby:

doing that? Can you do that as in? Can they see what other people are doing? Yeah, Could.

George:

Julia number one. See what all the other Julias are doing.

Robby:

If you set it up that way. Yes, if she has access, but can you see what your employees are doing?

George:

If I wanted to yeah, so why can't?

Robby:

this be set up like that. There's only one way that you would see your employees that's not transferable, and that's by sitting opposite them. Yeah, so the AI agent itself then do you need to train it per employee?

George:

No, why no? I'm just asking Because they're going to have different ways they operate. I don't mind if they have different ways of doing things.

Robby:

Your employees bend for the business, not the business doesn't bend for the employees.

George:

Yeah, but if they have a certain way of filing that works for them, for example, and they want to do it a certain way as long as they're getting the work done.

Robby:

I don't give a fuck yeah, but that's just a bad training. Yeah, you know. I mean, humans are hard to train yeah they're stuck in their ways, but you train this how you're real yeah you know what I mean.

Robby:

And if I, if I have 10 children and I train all of them the exact same way, they will all come out pretty similar, like you'll'll find. Like the family, they all do this or they all like this, whatever it is, they'll all come out fairly similar. Yes, you can argue oh, but this kid does this and this kid does that, and it's usually because there's a slight difference that you haven't paid attention to. In my opinion, I don't think humans are born with traits. Do you reckon humans are born with traits?

George:

No, they're given, they're picked up. Yeah, they're picked up?

Robby:

Yeah, they're picked up. Something you know what I mean? Like the, the traits are all picked up from things that we're not even aware of. Uh, we, we lack awareness around. Even though you can sit there and say I'm enlightened individual, it's like you still lack awareness, like you're not aware of everything. You cannot be aware of everything. You'd it'd be sensory overload to every single fucking thing that's happening. And this is like that. So you can train these. And if, technically, if I train one and this is my training regimen, so this is the and by training I mean feeding them information. Yeah, so I give them a bunch of information and they work off that. That becomes a knowledge base. If I go copy and paste that the next one should do the exact same thing, they don't sit there like. If I go copy and paste that the next one should do the exact same thing, they don't sit there like oh, but I like to do it like this.

Robby:

It's like no, you don't Do it this way, Then we don't. And it knows it's broken down into a series of steps.

George:

You know what I mean. Humans are complex. Yeah, so now I've got Greg. Greg working for me AI agent and you're actually building you've actually somewhat built this already and Greg's just going to call everyone that's going to come to our next event. Yeah, Say hey, Robbie, great to see you, Really excited to have you at the next success conference, such and such. Just wondering if you have any questions, whatever it might be. Yeah, so it's these agents actually calling every single person, yes, and getting the information You're actually building that for us as we speak.

George:

Yeah, that's pretty cool.

Robby:

Very cool.

George:

Yeah, now, instead of me paying my PA to do it for three days. This AI agent can now. Now, will it make one call at a time, or does it go out and make 30 calls or 50 calls at a time?

Robby:

So if you're going to do that, you've got to connect 50 phone numbers Right, which, unless you had a large amount.

George:

Yeah, 100,000 people you had to call.

Robby:

Another thing you need to factor in is like can it process that much information at once, if it's talking to 150 people at once, can it comprehend everything happening all at the same time and how powerful the models are? Because the models are getting more powerful.

George:

That's what I mean. Like I would imagine that this is at a point where it's getting to become limitless. Like what's the constraint? Is it electricity? Like, is it memory? Is it like what's the constraint?

Robby:

Yeah, it's that they aren't completely developed yet.

George:

Yeah.

Robby:

Yeah, that's what it is. They're developing over time, like the time, like they're getting smarter and smarter and smarter. And then, if you realize, there's different models. So if you go into chat gpt, and you'll see like, uh, you might have gpt 440, all right, or 40. And then there's 45, and there's 30, and there's 30 mini, there's 40 mini, and it's like everyone has slightly different, more capability. It's like, hey, man, I've trained this kid to be good at maths, but this, this one, she's a really good writer. Do you know what I mean? And it's like they're both great, but they're both more skilled in different ways.

Robby:

So, you can have different. They refer to it as the brain. So, like the three aspects the tools, the brain and the knowledge base you can have different brains for every single, and that's why you, you can have different brains for every single, and that's why you wouldn't have one doing everything.

George:

Yes, Now will it get to the point where you might have one doing everything. Yeah, maybe, yeah, that's right. Yeah, so that's cool that we've got that one getting built at the moment, because, thinking about, as you said, I don't have to pay it, it's getting it done in probably a lot quicker time than someone physically picking up a landline and calling and dialing the numbers and hey, how are you? Hey, how are you? Oh, I forgot to tell them that shit.

Robby:

It's also the cost.

Robby:

Yeah, do you know what I mean. So we use now we have a sales copilot and what that is is, before anyone gets on a sales call, we input someone's LinkedIn URL on their website and it goes and it scrapes their website, scrapes their LinkedIn URL and it tells us everything about the individual, the company, and then knows everything about our company. So how we fit in. So it'll be like hey, robbie, you're going to work with George. George's weak points in his business are this and this and this, and I think your company can help them really well from this angle. You know what I mean.

Robby:

It gives me some opening lines and this is what we've trained it on, so we can train it to be whatever we want. I can be like hey, you only give me three opening lines, give me five every single time, and so, okay, cool, hey, you mentioned this every single time, don't mention this anymore, okay, and we just continually refine it. Right, that's a very basic type model. But if I said to someone, someone on the team, hey, I'm going to meet with George from Pascon, can you go look at George's LinkedIn Profile? Yeah, go look at his website, and then I want you to provide me a full summary of how I can approach this call Like a one-page summary. How long would that take?

George:

Yeah, two, three hours, four hours, whatever, yeah.

Robby:

Time. How much is it going to cost?

George:

Yeah, exactly their hourly rate. Yeah, plus the opportunity cost of everything else they could do 30 bucks an hour.

Robby:

It's going to cost you a hundred bucks to get something like that done. It costs you about three cents and it happens instantly. So you get it within a minute maybe less like that, cause it goes. It looks at the whole. Linkedin comes back with this whole thing. I reckon it'd be close to a minute because it's got to scrape the data and do all that. But within a minute you've got this full summary. It's cost you three cents and all of a sudden now, like imagine you had someone dedicated to market research in your team. That person's gone. He's lost his job. Yeah, because now I've got someone who can do it a hundred times quicker for one hundredth the cost. How do you compete with that? You?

Robby:

don't yeah, so no cost or low cost yeah, low cost um no annual leave, no sick leave, no hr issues, no fair work letters uh, and it'll probably get to a point with very few mistakes. Yeah, well, it'll be less than once, and as it continues to develop it's just going to start to nail everything every single time.

George:

It may not be perfect. The first one.

Robby:

Yeah, but here's the best part. Here's the best part. So let's say you've got a really complex thing that you do. Let's just say, for example, tenders, quotes, tenders, estimates. There's a lot of detail in an estimate, right? So then you notice, fuck man, it keeps making this error, or it keeps making that error, or one in every five there seems to be an error. Okay, so let's build an AI agent that can check this for them. Yeah, it's like let's build something that can then refine it and is taught to train this and tell it when it makes a mistake. And then all of a sudden, you've got Greg and Julia talking and Greg's training Julia, because obviously men are better. Do you know what I mean?

George:

Yeah.

Robby:

And it's like, okay, cool Now, that thing I had to do manually I don't have to do anymore. It's like it's already getting done.

George:

Because that's where I saw or maybe you mentioned it like that's where I saw a huge inefficiency in my business was the tendering aspect of things, and you know there's even human error to take into consideration.

Robby:

Yeah.

George:

Have I made a mistake on doing a takeoff or did my team make a mistake on doing this or doing this, whereas this will be able to look at the plans, interpret them and, in 10 minutes, give you. Okay, george, this is going to cost $6.4 million, yeah $6.4 million $6,483,913.

Robby:

483,913 cents, yeah, and probably be more accurate than, or at least get to the point where it will be more accurate than.

George:

And that's yeah, and that's an agent. That man you build that it's worth millions. Yeah, to a person building that, like as into a company that's worth so much money Because all of a sudden, what used to take you six weeks time.

Robby:

This is where you're getting wrong. It's not an agent. It might be 40 agents.

George:

Yeah, whatever it is you know what I mean. So would you have one that does the takeoff, one that does the money, one? That does the whatever, it is Probably the proposal.

Robby:

Yeah, and then one that you give the information to at the start and say, go, sort it out. And then he goes and speaks to every single one and then comes back to you 20 minutes later and says yeah. And then you're like what the fuck are you doing? And you look at it and you start flipping and you go and you say what the fuck? You guys want your jobs because I'll replace you on christmas yeah, you want your christmas bonus.

Robby:

Do you want to go home and see your families? Because I will replace you very fast.

George:

Yeah, and that's the sort of shit that's so powerful there. You know that's from me thinking about it. You nailed that. For me that's worth its weight in gold, especially like if there's a level of nervous. I'd be checking it at the beginning anyway, as in probably doing it manually myself and getting them to do it to see how accurate it gets or how close it gets.

Robby:

And then you're like it's off. Two nights You're off.

George:

Yeah, two nights ago, no, yesterday I submitted a tender and for me to put the actual, so I did all the numbers, had everything there, it was all good to go, and then putting the actual proposal together took me throughout the day, let's say a couple of hours. All right, I know for a fact that you train an agent to do that. It'd be done in like three minutes or in 30 seconds.

Robby:

This is where people will get let down. It's like most people don't know what they want. Yeah, and it's like they're like you like put the proposal together and then they kind of work at it as they go and they're like, yeah, this is good enough, whatever, you know what I mean. You need to have things templated, defined, that's right.

George:

And I've kind of got that already. So I've got a lot of templates and I can say this is an example of the proposal that I give. Yeah, these infamous all this thing, all this, all this blah, blah, blah blah. It needs to include all these things, and then obviously I would still look through it because it might be specifics or things that I need to update or whatever it might be. But ultimately, instead of that taking me two hours or three hours to put together, it takes 10.

Robby:

It might take you 20 minutes.

George:

Yeah, you know what I mean. It's a much simpler process and it saves me so much time. Not only has it done a six-week job for me in half an hour, it's then done the three-hour task for me in five minutes and taken me 20 to do.

Robby:

Yeah and it's like the same way you would look over it when you gave it to a new employee, and that's the exact same thing and you'd still say let me check it. That's right, Let me just make sure that this was right and you'd see what they did and you'd be happy with it within your company's Companies' guidelines. The exact same thing Until you get to the point where you're like I know they're good, I don't have to check this anymore, that's right. That's right, they are good. The same way you build that trust with an employee.

Robby:

You're going to build it the same way here. The same way they're going to understand how you communicate. It's going to be the same thing here. The AI is going to learn and it's going to be like I'm going to double check this because George gets angry. Because George gets angry when there's an error on the last bit and it takes that in. It's got the ability to reason, which is thinking. You know what I mean and this is something we haven't dealt with before.

George:

Yeah, exactly.

Robby:

It's going to change everything, dude.

George:

Yeah, man, it's so cool. Do you know what would be great? It's like when we double back on this conversation in 12 months' time, you know. And then we've got agents in the business, You've got things that are happening or you see where the space is and everyone knows what an agent is and it's like a normal part of life. And then you see everyday people using agents, even like a mum or a dad to order their coffee you know in the morning, yeah, but that's going to happen.

Robby:

This is what I think people aren't understanding. It's going to happen. This is what I think people aren't understanding. It's gonna get to the point where it's. Imagine. At one point people were like wait you, you connect the old thing, you plug it into the wall and you connect and you dial something and it makes these weird noises and you connect and you can talk to other people somewhere else. Yeah, like that's bizarre, like how?

George:

do you know who's on the other side? Yeah, how do you know?

Robby:

who's on the other side. That's fucking insane.

George:

Do you?

Robby:

know what I mean. And now it's like you fucking take Wi-Fi off a kid, he turns into a crackhead. Do you know what I mean? They fucking lose their shit. We can't live without the internet. We've become so dependent on it. This is that People are going to become so dependent on it, Like so, so, so dependent on it because this is so powerful. It's just going to become a way of life and we're not going to remember what life was like. Can you imagine not being able to find something out instantly?

George:

Yeah.

Robby:

Like can you think about that? There was people who went their whole life without being able to answer a question.

George:

Yeah, their whole life, which is why it was, like you know, even at school they taught the capital of Brazil, you know, and all these things you're taught, these useless things Like, why would we need to learn all of these things now, when I can just pick up my phone and Google it?

Robby:

Yeah, if I want to know what the thing is, I can learn.

George:

Like it's not an issue anymore.

Robby:

Because the way we're living is going to change. Like I think the next thing is going to be about leverage, like how do I build systems that I can leverage in my life? Do you know what I mean? And there's a whole bunch of doomsday theories where they think humans are going to become irrelevant, like we're going to be the less superior species. But that isn't going to happen. We're going to be the less superior. This is going to Sam Altman, the CEO of OpenAI, which had JT turned around and said my kids are never going to be smarter than AI. Yeah, he's like I have to accept that they're never going to be smarter than it Ever, ever they're going to grow up in. He's like is there downsides to that? Yes, he is, but there's also many upsides.

George:

Yeah, it's like man. It's how you take it.

Robby:

Yeah, but it's happening, it's happening.

George:

Yeah, like man, you could talk, you could have a massive conversation about this for a very long time, this for a very long time. You know what?

Robby:

I mean.

George:

Yeah, very long time and it's great. It's exciting because it's the first of many conversations. I think we're going to have.

Robby:

We'll keep everyone posted with how this is coming along.

George:

That's it. But the only way you'll actually know about the status of this conversation is if you are subscribed to the podcast. So if you haven't done so already, definitely do it, and then tell all your friends about it, because that's how we can reach many people, have interesting and in-depth conversations day in, day out, week in, week out.

Robby:

Week in, week out, every Monday New episode landing.

George:

You know where to find us. That's it All right, guys. Thank you so much. I hope we've been. I hope we've just tickled something inside of you. Yeah, I hope you catch. I hope you catch the bug.

Robby:

Yeah, and I hope you go and you do something with this and you say hey, let me just look into let me just Google. These guys are weird but, let me look into it a little bit to see what's coming, because I think most people have no idea what's coming. Dude, like no fucking idea.

George:

And on that bombshell, I hope you all have a million-dollar day and we will see you next week. Thank you so much, guys. Thanks everybody. See you yeah.

Robby:

Camera just died.

George:

Huh, She'll be right, we've got the front camera.

Robby:

Yeah.

George:

Processing. I'm going to fuck the audio you up.

People on this episode

Contributor

George Passas

Host
Contributor

Robby Choucair

Host