Show Notes
Arlo Zephyr is Scott's first AI agent. Built on OpenClaw. Named by Scott's wife. Last name comes from Zephyr Hills, Florida.
Tool vs. Team Member: Most people think of AI as a tool—ready to use like a hammer. Scott treats Arlo as a team member. That means onboarding, training, watching limitations, and bringing in humans when needed. "It's going to be terrible in the beginning."
The 5 Use Cases:
1. Morning Brief Arlo aggregates revenue from multiple sources, delivers metrics, and stops the obsession of checking email for every sale notification. "Disney isn't getting an email every time someone swipes a card. Revenue is expected."
2. Email Distribution Arlo pulls data, compiles newsletters, formats HTML, and schedules emails. Freed up ~10 hours/week for the team to do other things.
3. Customer Account Analysis (Land Moto) Arlo reviews customer listings, identifies improvement opportunities, and feeds research to humans who reach out. This is a new capability—work that wasn't being done before.
4. Phone Calls (Testing) Early experiments with inbound and outbound calls. Results are mixed. "I called a restaurant, talked to an AI, and asked for a human." It will get better, but it's not there yet.
5. Cold Email Outreach (B2B) Arlo identifies 20-30 leads/day, does research, and explains why each is a fit. Humans review and send. The temptation is to let it run wild—don't. Human in the loop.
Key Insight: "Capability" is the word. Adding Arlo is like adding a team member—you gain new capabilities the organization didn't have before.
Human in the Loop: Repeated throughout. AI makes things up (hallucinations). Humans must review before sending.
Connection to SCALE: "I scoped it. I clarified the flow. I built one use case, watched it, and added more when comfortable." The framework enabled Arlo.
Your Action: Think about one task you could use an AI agent for. Map it out with SCALE. Then figure out how to turn it into an automated agent.
Bonus — Edward:
Scott's new daily micro-parables on Substack. Edward is a composite of every business owner. Follow at scotttodd.net/blog.
Got a business question? Ask Scott here: scotttodd.net/ask
📜 Full Transcript (Click to expand)
Scott Todd (00:00)
I have an AI agent. His name is Arlo, and he works across my businesses. And today I'm going to share five use cases and how I'm using them and what he has freed up for my team to do instead. Welcome to Fix My Business. I'm your host, Scott Todd. I've built multiple seven-figure businesses after leaving my Fortune 300 VP position. And this channel is dedicated to helping you.
Do the same. Now, who is Arlo? What is he? And how did he come into existence? Well, look, there's no it's no secret that in 2026 AI agents are blowing up. Okay. AI agents are everywhere. Everybody's talking about the AI agents. And it's almost like a race to see how many AI agents you can have. And I have a couple of them.
But I want to tell you about Arlo because he is my first and most loyal AI agent if that thing exists. Okay. So here is here is Arlo. Arlo is built on that open claw framework. And then if you're like, I don't even know what Scott's saying, it's okay. I just don't want you to get caught up on the technology. I want you to really understand kind of what he's doing and what he's freeing up. That's the more important message here. Okay. So don't worry about the tech. The tech can, we can find people to help with the tech.
If that's not you, and that's okay. So how did his name come into existence? Well, his full name is Arlo Zephyr. Interesting name. A Z. Arlo Zephyr is his initials. And the way that Arlo came into existence is like I needed to give him a name. I needed some way of calling to calling him, referring to him, to give him some for sort of like some image in my brain of how I could work with him. And really his name was ⁓ just.
Asking my wife, like, hey, I got this AI agent. What should we name them? We brainstormed some names and she liked Arlo. And I was like, Arlo, it is. So that's how Arlo came into existence. I wanted to give them a last name. Why? I don't know. I don't know. It I did look, I grew up in the I grew up in the late 70s. I grew up in the 70s, 80s, cabbage patch dolls. My sister had a cabbage patch doll. You could name the cabbage patch dolls. ⁓ is this the new cabbage patch doll? I don't know.
So I gave him last name. His last name is Zephyr. And it's just basically my office is in the city of Zephyr Hills, Florida. And guess what? Zephyr. I thought it was kind of cool. So isn't that cool? That's how his name came into existence. Now, one of the things that I think that separate people that are using AI agents is in the way in which they think about their agent. Because it's, I think the the way that you think about it.
Changes the way that you use it. So let me give an example. I know that some people just look at AI and they look at, you know, their AI agent and they think, this is just another tool that I use. And that's fine. I like to think of them as a team member. And there's not a right or wrong way to think about an AI agent. But when I think about them as a team member, now I'm thinking and approaching.
His onboarding in a different way. See, to me, I think that most people, when they think about something being a tool, they think that that tool is already ready to use. So look, if I pick up the hammer, because that's a tool, I'm gonna pick up the hammer and I'm gonna use the hammer. I don't need to train the hammer to do the hammer's job. Okay, so tools are just things that you plug into. But when you think about an AI agent,
More as a team member, where there's going to be an onboarding period and it's going to be terrible in the beginning. And it's going to just be just not what you expected. And it's not going to be any fun. A team member gives you the right mindset in order to onboard them correctly. And the way that I have onboarded all of the agents that are kind of in my company is I treat them like a new employee, I feed them information.
I teach them things and I kind of ease into this. So I try to write material processes that can kind of help them to do the work. I watch it where their limitations are. I either modify it or I pull in a human in the loop, important piece, when we're talking about the AI agent not being able to do some of the things that I want it to do.
Technology and automation in a business is this blending of humans and machine. Let the machines do what it can do best. Let the technology do what it can do best. So, what are the five use cases? Simple. Number one use case. And I think that everybody that has an AI agent probably uses this one, and I'm no different, is my morning brief. Now, here's what used to happen is I used to be like a lot of you, where
I started, you know, you get these credit card notifications like you've made a sale and you're like, Yes. And then all of a sudden you get more and more and more of them. And it's it's a fantastic day. And then guess what? Then you start obsessing in your email looking for that hit. Like, my gosh, ⁓ did I make any money yet? Did I make any sales? How many of you are still looking at your email constantly? And then you're disappointed when you don't have a notification that you made a sale or collected some money.
It's a hit. And I got tired of that. So what I did was I shut down all those notifications. I was like, you know what?
Big companies, they're not getting an email every time they make a credit card swipe. Can you imagine like Disney? I I go to Disney and I think, my gosh, every time I swipe that credit card, if an executive was getting an email saying Scott just made a ⁓ or this food cart just made a per ⁓ a sale, my gosh. They'd be drowning in the email. They couldn't do it. It's expected. If you're running a business, it's expected. Revenue is expected.
So, what I did was I stopped getting those notifications and I would get the daily wrap up. And that's fine. But then I had multiple sources, and I'm trying to do the math in my head and trying to pull all this together. So, what I asked Arlo to do is to get give me a daily briefing that shows the sources of my revenue, what business is generated and what revenue. Also, some metrics that are important to me. Okay, so there's some business metrics that are important to me, there are vanity metrics.
They move the needle in no way, but instead of me obsessing over it and looking at it throughout the day, I just said, you know what, I'm gonna get it tomorrow. And he he delivers me the vanity metrics that I hate looking at anyway. So it's this morning brief, but Arlo also has become this expert in this material, knowing what their credit card transactions are.
Being able to answer my question. See, that morning brief kind of gave them access to that that data. So if I have questions about, hey, did somebody pay? or hey, what about this? Arlo will surprise me now if he's working on other things because he knows that he has access to the credit card data and he will come back to me and say, by the way, this is wrong because or this is right because these people paid. Wow.
Okay, that's kind of cool. Thank you, Arlo. Use case number two is email distribution. And I have a couple of what I would call email distributions where I want emails to go out on a specific time. And the way that it used to work is I would have an employee who would go into, let's say, kit.com or activecampaign.com, where we manage our contacts and they would schedule the human would.
Prepare the email and schedule it. Okay. So that meant that I had human beings that would sit down, craft these emails, they would then take them, go into the service provider, manually schedule it all out every single week. And it was time. I mean, with all of the emails that we were sending, it was time.
It was probably around 10 hours a week minimum. And that's just kind of the writing of the email and scheduling them. Okay. And some of these emails, especially with newsletters, it's kind of just pooling data that's already exists into one place, putting it into HTML and making it look pretty, and then sending it out. Okay. Well, Arlo's able to do that. Arlo's able to look at the sources that we would normally pull this data from that compiles these emails, pull them in together and
Put them into one email that we can send out and schedule it. And that's what email Arlo has done with our email. So he's freed up about 10 hours of week a week for people, for my team to go do other things. And that's an important piece because when we're freeing up time, we want to redeploy that time into something that will generate more revenue. Use case number three is basically for.
my company Land Moto, he will go in on a regular basis and he will look at customer accounts. He will look at the results that they're generating from the listings that they have. He will determine ways in which to improve those listings. And then he will send it to you to a human who will reach out and say, Hey, I have some ideas for you. I have some ideas that can help you write better ads. I have some ideas that that ⁓ might might improve your exposure.
But what he's doing is he's feeding into the human research that the human would have done. It was honestly work that we weren't doing. So this is a newfound capability that the company has because of Arlo's existence. And I think that that's an important word. I want to stress the word that I just used there, capability. Because
Whenever you add a new employee to your team, you're hoping to pick up new capabilities. I mean, you bring your own capabilities to the organization. Everybody that works on your team has their capabilities. And the more capabilities that you stack into each other, ultimately you build a stronger organization. So the mere fact that Arlo is now able to go out and to identify accounts that basically might need some TLC, that's a newfound capability.
That we can provide a better service for. So that is one of the things, but if he works with a human. He's not. Listen, I've thought about having Arlo, just being honest, I thought about having Arlo get on the phone and call you. Call our customers and say, hey, here's a way to improve. I don't know if that would freak you out or not. Maybe I should try it. But that does lead me into use case number four, which is phone calls. I'm messing around with Arlo on phone calls. And
He's okay. He's not bad. At the same time, it can be a little frustrating talking to an AI. I mean, I called to ⁓ a restaurant the other day to place an order and I wanted something specific. I wanted to kind of let them know ⁓ of a food allergy or a special request, not a food allergy, a special request. So I called in and I asked, like, place it to go order, please. The next thing I know, I'm talking to an AI agent.
And I'm like, this is not gonna work. So then I ask for a human. Well, then I get a human. Okay. So we all know that AI is not necessarily the greatest experience. It cuts you off. It doesn't have that human intuition yet. I'm sure that it will come. But you know, one of the use cases that I have been testing out and trying is ⁓ anything that has to do with making phone calls or receiving phone calls. And ultimately, you know, back in the back in the day.
I remember when before the internet, way back then before the internet, you would drive down, you know, streets and look for homes for sale, and they would have phone numbers on the signs. And the phone numbers on the signs were specific to that house, and you'd call and it would give you all of the data around the house, all of the listing data. Now, look, the internet has replaced that. But what Arlo is testing for me is ways in which we can kind of take some of these inbound calls from customers.
To relieve the human in the loop. Is there things that employ that Arlo can deal with, like ⁓ simple requests, such as, hey, I'd like to, ⁓ you know, change a credit card that I have on file. No problem. Arlo can capture that information, even if he doesn't tie into the systems yet. He can capture that information and then he can give it to the human in the loop.
But look, this is early. I don't love the results, but it's early. Okay. It will get better. I have no doubt it will get better. The fifth use case that I have Arlo using for me in one of my businesses is essentially cold email outreach. Now, this business is a ⁓ B2B company. And what Arlo is doing is Arlo is identifying 20 to 30 potential leads a day.
And what he's doing is he's doing the research to explain why he believes this is a fantastic connection with my company. And he's feeding the leads, he emails the leads to my sales team, who then takes that, reviews it, and then he turns around and they will send out the email.
Sometimes they'll modify what Arlo says. Sometimes it's like good and they send it. But the human in the loop is the review. And that is a temptation. The temptation is to just turn these AI agents loose and let them go crazy. But I want to caution you. I want you to remember you have to have the human in the loop. It's on, it's cutting edge, it's modern. And they have a way of making things up sometimes called hallucinations.
And you want to make sure that your humans are making sure that that is not happening. I do know a real estate investor who has successfully taken their AI agent, trained it in sales through text, and it will text back and forth with prospective customers. The last I heard, there was a number of sales, I think like five or six sales that the AI agent was able to make, but the human was approving every single email or text before it would go back.
So the human in the loop was looking at the the email, the text, and making the decision to send it or not to send it. Okay. So here's the reality: like all of this has shifted our capacity, right? Like the capability, we have new capabilities within the company, but we also have new capacity. And when you have more capabilities and more capacity, guess what? You have a much different business. And what's happened here is that.
Arlo has created a situation to where we have freed up time, but we're also able and to do things that we were never able to do before. And so the whole point of this episode is really to go back and say, well, how do we put something like this in place? And over the last few weeks, we have looked at this automation and process framework that I use called the scale framework. And scale ultimately comes back to sitting down.
looking at the pain points within the business, identifying a way in which you can automate those, but you first have to revive or ⁓ revise the process first. And then what you want to do is then you can go back in and build this process in an automated way. So the way that I did this was I would scope out a work, I scoped it out, scale a scope. I scoped out exactly what I wanted this process
process to do. I clarified the flow, which means that I literally mapped out every step. I'm a big believer in mapping out the steps before you build a process. It's like building up like it's like the NFL when they build a play and they map out what's going to happen or who's going to be in what position. That's the same way I think about building with scale. Okay. And I did not try to build all five of these at one time. I built one use case
I watched it, we we ran it, looked at it. When it when we were comfortable with it, we added more. The scale process that we've spent weeks talking about, all of the traps, basically capturing the friction points, the build session from the last episode, all of these now give me the strength to move fast to deploy new technologies. And that's the thing is like Arlo is still evolving. And you'll hear more about him.
On this channel because he's going to be more and more larger piece of my business. I don't think of him as a personal assistant. I think of him as a valuable team member. A team member that can free up and work with my other teams, my humans, to give us greater capacity and greater capability. Now, here's the action: I want you to think about one task in your business that you could use an AI agent for.
What can you do? How can you map that out? And map it out with scale and then figure out how you can turn that into an automated agent. Okay. Okay. Now, before I close, I want to leave you with one thing. Over on my Substack. And if you if you're not reading that or if you're not following me over there, please consider doing it. It's at Scott, just go to Scott Todd.net. Up at the very top, it says blog.
Click on that and it will take you right there. And I have started ⁓ a daily, a daily ⁓ micro parable where I introduce this character called Edward. And Edward is a composite of every business owner I've ever dealt with and the challenges that you are probably facing. And every day there's some lesson there that I'm writing so that you can look at your business in a whole new way.
And it's that challenge of that delicate balance between knowing what to do and then the emotional decisions that we all carry as business owners. And I'd love to have you ⁓ along with Edward's journey because it's also your journey. All right. So head over to Scott Todd dot net, just click blog at the upper corner up there, up at the very top, and I will see you in our next episode.
I have an AI agent. His name is Arlo, and he works across my businesses. And today I'm going to share five use cases and how I'm using them and what he has freed up for my team to do instead. Welcome to Fix My Business. I'm your host, Scott Todd. I've built multiple seven-figure businesses after leaving my Fortune 300 VP position. And this channel is dedicated to helping you.
Do the same. Now, who is Arlo? What is he? And how did he come into existence? Well, look, there's no it's no secret that in 2026 AI agents are blowing up. Okay. AI agents are everywhere. Everybody's talking about the AI agents. And it's almost like a race to see how many AI agents you can have. And I have a couple of them.
But I want to tell you about Arlo because he is my first and most loyal AI agent if that thing exists. Okay. So here is here is Arlo. Arlo is built on that open claw framework. And then if you're like, I don't even know what Scott's saying, it's okay. I just don't want you to get caught up on the technology. I want you to really understand kind of what he's doing and what he's freeing up. That's the more important message here. Okay. So don't worry about the tech. The tech can, we can find people to help with the tech.
If that's not you, and that's okay. So how did his name come into existence? Well, his full name is Arlo Zephyr. Interesting name. A Z. Arlo Zephyr is his initials. And the way that Arlo came into existence is like I needed to give him a name. I needed some way of calling to calling him, referring to him, to give him some for sort of like some image in my brain of how I could work with him. And really his name was ⁓ just.
Asking my wife, like, hey, I got this AI agent. What should we name them? We brainstormed some names and she liked Arlo. And I was like, Arlo, it is. So that's how Arlo came into existence. I wanted to give them a last name. Why? I don't know. I don't know. It I did look, I grew up in the I grew up in the late 70s. I grew up in the 70s, 80s, cabbage patch dolls. My sister had a cabbage patch doll. You could name the cabbage patch dolls. ⁓ is this the new cabbage patch doll? I don't know.
So I gave him last name. His last name is Zephyr. And it's just basically my office is in the city of Zephyr Hills, Florida. And guess what? Zephyr. I thought it was kind of cool. So isn't that cool? That's how his name came into existence. Now, one of the things that I think that separate people that are using AI agents is in the way in which they think about their agent. Because it's, I think the the way that you think about it.
Changes the way that you use it. So let me give an example. I know that some people just look at AI and they look at, you know, their AI agent and they think, this is just another tool that I use. And that's fine. I like to think of them as a team member. And there's not a right or wrong way to think about an AI agent. But when I think about them as a team member, now I'm thinking and approaching.
His onboarding in a different way. See, to me, I think that most people, when they think about something being a tool, they think that that tool is already ready to use. So look, if I pick up the hammer, because that's a tool, I'm gonna pick up the hammer and I'm gonna use the hammer. I don't need to train the hammer to do the hammer's job. Okay, so tools are just things that you plug into. But when you think about an AI agent,
More as a team member, where there's going to be an onboarding period and it's going to be terrible in the beginning. And it's going to just be just not what you expected. And it's not going to be any fun. A team member gives you the right mindset in order to onboard them correctly. And the way that I have onboarded all of the agents that are kind of in my company is I treat them like a new employee, I feed them information.
I teach them things and I kind of ease into this. So I try to write material processes that can kind of help them to do the work. I watch it where their limitations are. I either modify it or I pull in a human in the loop, important piece, when we're talking about the AI agent not being able to do some of the things that I want it to do.
Technology and automation in a business is this blending of humans and machine. Let the machines do what it can do best. Let the technology do what it can do best. So, what are the five use cases? Simple. Number one use case. And I think that everybody that has an AI agent probably uses this one, and I'm no different, is my morning brief. Now, here's what used to happen is I used to be like a lot of you, where
I started, you know, you get these credit card notifications like you've made a sale and you're like, Yes. And then all of a sudden you get more and more and more of them. And it's it's a fantastic day. And then guess what? Then you start obsessing in your email looking for that hit. Like, my gosh, ⁓ did I make any money yet? Did I make any sales? How many of you are still looking at your email constantly? And then you're disappointed when you don't have a notification that you made a sale or collected some money.
It's a hit. And I got tired of that. So what I did was I shut down all those notifications. I was like, you know what?
Big companies, they're not getting an email every time they make a credit card swipe. Can you imagine like Disney? I I go to Disney and I think, my gosh, every time I swipe that credit card, if an executive was getting an email saying Scott just made a ⁓ or this food cart just made a per ⁓ a sale, my gosh. They'd be drowning in the email. They couldn't do it. It's expected. If you're running a business, it's expected. Revenue is expected.
So, what I did was I stopped getting those notifications and I would get the daily wrap up. And that's fine. But then I had multiple sources, and I'm trying to do the math in my head and trying to pull all this together. So, what I asked Arlo to do is to get give me a daily briefing that shows the sources of my revenue, what business is generated and what revenue. Also, some metrics that are important to me. Okay, so there's some business metrics that are important to me, there are vanity metrics.
They move the needle in no way, but instead of me obsessing over it and looking at it throughout the day, I just said, you know what, I'm gonna get it tomorrow. And he he delivers me the vanity metrics that I hate looking at anyway. So it's this morning brief, but Arlo also has become this expert in this material, knowing what their credit card transactions are.
Being able to answer my question. See, that morning brief kind of gave them access to that that data. So if I have questions about, hey, did somebody pay? or hey, what about this? Arlo will surprise me now if he's working on other things because he knows that he has access to the credit card data and he will come back to me and say, by the way, this is wrong because or this is right because these people paid. Wow.
Okay, that's kind of cool. Thank you, Arlo. Use case number two is email distribution. And I have a couple of what I would call email distributions where I want emails to go out on a specific time. And the way that it used to work is I would have an employee who would go into, let's say, kit.com or activecampaign.com, where we manage our contacts and they would schedule the human would.
Prepare the email and schedule it. Okay. So that meant that I had human beings that would sit down, craft these emails, they would then take them, go into the service provider, manually schedule it all out every single week. And it was time. I mean, with all of the emails that we were sending, it was time.
It was probably around 10 hours a week minimum. And that's just kind of the writing of the email and scheduling them. Okay. And some of these emails, especially with newsletters, it's kind of just pooling data that's already exists into one place, putting it into HTML and making it look pretty, and then sending it out. Okay. Well, Arlo's able to do that. Arlo's able to look at the sources that we would normally pull this data from that compiles these emails, pull them in together and
Put them into one email that we can send out and schedule it. And that's what email Arlo has done with our email. So he's freed up about 10 hours of week a week for people, for my team to go do other things. And that's an important piece because when we're freeing up time, we want to redeploy that time into something that will generate more revenue. Use case number three is basically for.
my company Land Moto, he will go in on a regular basis and he will look at customer accounts. He will look at the results that they're generating from the listings that they have. He will determine ways in which to improve those listings. And then he will send it to you to a human who will reach out and say, Hey, I have some ideas for you. I have some ideas that can help you write better ads. I have some ideas that that ⁓ might might improve your exposure.
But what he's doing is he's feeding into the human research that the human would have done. It was honestly work that we weren't doing. So this is a newfound capability that the company has because of Arlo's existence. And I think that that's an important word. I want to stress the word that I just used there, capability. Because
Whenever you add a new employee to your team, you're hoping to pick up new capabilities. I mean, you bring your own capabilities to the organization. Everybody that works on your team has their capabilities. And the more capabilities that you stack into each other, ultimately you build a stronger organization. So the mere fact that Arlo is now able to go out and to identify accounts that basically might need some TLC, that's a newfound capability.
That we can provide a better service for. So that is one of the things, but if he works with a human. He's not. Listen, I've thought about having Arlo, just being honest, I thought about having Arlo get on the phone and call you. Call our customers and say, hey, here's a way to improve. I don't know if that would freak you out or not. Maybe I should try it. But that does lead me into use case number four, which is phone calls. I'm messing around with Arlo on phone calls. And
He's okay. He's not bad. At the same time, it can be a little frustrating talking to an AI. I mean, I called to ⁓ a restaurant the other day to place an order and I wanted something specific. I wanted to kind of let them know ⁓ of a food allergy or a special request, not a food allergy, a special request. So I called in and I asked, like, place it to go order, please. The next thing I know, I'm talking to an AI agent.
And I'm like, this is not gonna work. So then I ask for a human. Well, then I get a human. Okay. So we all know that AI is not necessarily the greatest experience. It cuts you off. It doesn't have that human intuition yet. I'm sure that it will come. But you know, one of the use cases that I have been testing out and trying is ⁓ anything that has to do with making phone calls or receiving phone calls. And ultimately, you know, back in the back in the day.
I remember when before the internet, way back then before the internet, you would drive down, you know, streets and look for homes for sale, and they would have phone numbers on the signs. And the phone numbers on the signs were specific to that house, and you'd call and it would give you all of the data around the house, all of the listing data. Now, look, the internet has replaced that. But what Arlo is testing for me is ways in which we can kind of take some of these inbound calls from customers.
To relieve the human in the loop. Is there things that employ that Arlo can deal with, like ⁓ simple requests, such as, hey, I'd like to, ⁓ you know, change a credit card that I have on file. No problem. Arlo can capture that information, even if he doesn't tie into the systems yet. He can capture that information and then he can give it to the human in the loop.
But look, this is early. I don't love the results, but it's early. Okay. It will get better. I have no doubt it will get better. The fifth use case that I have Arlo using for me in one of my businesses is essentially cold email outreach. Now, this business is a ⁓ B2B company. And what Arlo is doing is Arlo is identifying 20 to 30 potential leads a day.
And what he's doing is he's doing the research to explain why he believes this is a fantastic connection with my company. And he's feeding the leads, he emails the leads to my sales team, who then takes that, reviews it, and then he turns around and they will send out the email.
Sometimes they'll modify what Arlo says. Sometimes it's like good and they send it. But the human in the loop is the review. And that is a temptation. The temptation is to just turn these AI agents loose and let them go crazy. But I want to caution you. I want you to remember you have to have the human in the loop. It's on, it's cutting edge, it's modern. And they have a way of making things up sometimes called hallucinations.
And you want to make sure that your humans are making sure that that is not happening. I do know a real estate investor who has successfully taken their AI agent, trained it in sales through text, and it will text back and forth with prospective customers. The last I heard, there was a number of sales, I think like five or six sales that the AI agent was able to make, but the human was approving every single email or text before it would go back.
So the human in the loop was looking at the the email, the text, and making the decision to send it or not to send it. Okay. So here's the reality: like all of this has shifted our capacity, right? Like the capability, we have new capabilities within the company, but we also have new capacity. And when you have more capabilities and more capacity, guess what? You have a much different business. And what's happened here is that.
Arlo has created a situation to where we have freed up time, but we're also able and to do things that we were never able to do before. And so the whole point of this episode is really to go back and say, well, how do we put something like this in place? And over the last few weeks, we have looked at this automation and process framework that I use called the scale framework. And scale ultimately comes back to sitting down.
looking at the pain points within the business, identifying a way in which you can automate those, but you first have to revive or ⁓ revise the process first. And then what you want to do is then you can go back in and build this process in an automated way. So the way that I did this was I would scope out a work, I scoped it out, scale a scope. I scoped out exactly what I wanted this process
process to do. I clarified the flow, which means that I literally mapped out every step. I'm a big believer in mapping out the steps before you build a process. It's like building up like it's like the NFL when they build a play and they map out what's going to happen or who's going to be in what position. That's the same way I think about building with scale. Okay. And I did not try to build all five of these at one time. I built one use case
I watched it, we we ran it, looked at it. When it when we were comfortable with it, we added more. The scale process that we've spent weeks talking about, all of the traps, basically capturing the friction points, the build session from the last episode, all of these now give me the strength to move fast to deploy new technologies. And that's the thing is like Arlo is still evolving. And you'll hear more about him.
On this channel because he's going to be more and more larger piece of my business. I don't think of him as a personal assistant. I think of him as a valuable team member. A team member that can free up and work with my other teams, my humans, to give us greater capacity and greater capability. Now, here's the action: I want you to think about one task in your business that you could use an AI agent for.
What can you do? How can you map that out? And map it out with scale and then figure out how you can turn that into an automated agent. Okay. Okay. Now, before I close, I want to leave you with one thing. Over on my Substack. And if you if you're not reading that or if you're not following me over there, please consider doing it. It's at Scott, just go to Scott Todd.net. Up at the very top, it says blog.
Click on that and it will take you right there. And I have started ⁓ a daily, a daily ⁓ micro parable where I introduce this character called Edward. And Edward is a composite of every business owner I've ever dealt with and the challenges that you are probably facing. And every day there's some lesson there that I'm writing so that you can look at your business in a whole new way.
And it's that challenge of that delicate balance between knowing what to do and then the emotional decisions that we all carry as business owners. And I'd love to have you ⁓ along with Edward's journey because it's also your journey. All right. So head over to Scott Todd dot net, just click blog at the upper corner up there, up at the very top, and I will see you in our next episode.