Want Ai Agents in your business? watch this...
In this instructional video, Rory Ridgers explains what AI agents are, how they function, and why business owners should consider them over human employees. He defines AI agents as LLM‑driven software that takes inputs, follows predefined prompts, utilizes external tools, and produces outputs. Using a lead‑qualification example, he illustrates the agent architecture (input → LLM → tools → prompt → output). He then contrasts AI agents with human staff, highlighting advantages such as lower long‑term cost, speed, 24/7 availability, and rapid training, while noting limitations in flexibility, creativity, and nuanced decision‑making. To help viewers identify suitable tasks, Rory introduces the ROUTINE framework (Repetitive, Objective‑based, Uncreative, Time‑consuming, Input‑heavy, Numeric, Error‑prone) and breaks down each criterion. He enumerates concrete use‑cases—lead qualification, customer support, inbox management, market research, invoice processing, internal help desk, data entry/cleanup, content repurposing, calendar coordination, proposal drafting—and explains why they fit the framework. The video also outlines a typical agency workflow for building an AI agent system (discovery call, process mapping, prototype, fine‑tuning, deployment). Finally, Rory warns against common pitfalls: automating tasks that need human input, skipping basic automation, chasing hype over ROI, and under‑utilizing agents due to knowledge gaps. He ends with a call to action to book a free consultation.
Chapters
What's up guys, in this video I wanted to go over what actually are AI agents and how and when do we use them in a business. Just quickly before we get into it, if you haven't seen me before, my name is Rory Ridgers and I help business owners to integrate AI agent systems into their businesses. So if that's something you might be interested in, I've left the link to my website in the description. You can go ahead and book up a free call, we can talk about some ways to integrate AI into your business. So I've put together a little presentation that I want to run you through, so without further ado, let's get into the video. Okay guys, so first of all, we're going to start off with what actually are AI agents, because you guys might be completely new to this space. Some of you might know a little bit about it. Some of you might be a little bit more advanced, but we're gonna begin with what actually are AI agents for those who don't know. So the way I like to describe AI agents are AI agents are AI-powered pieces of software that can make decisions based on an input and use tools that have been given to get a desired output. So if that doesn't make much sense, I've put together a little diagram so you can actually visualize what I mean by that. So basically we have an input here that's going to go into an AI model or an LLM, which in this case can be something like ChatGPT. I use ChatGPT to power, uh, majority of my AI agents. So an input coming into ChatGPT, and then you can see down here we've given it some tools that allow it to do extra stuff outside of the normal realms of ChatGPT. This will also have a prompt, which is just instructions on basically what to do, how to do it, what tools it has down here. And then once it's done that, it'll give us some form of output. So usually when you're just chatting to ChatGPT, I'm sure all of you that are watching this video have chatted to ChatGPT at some point. Usually it would just be an input, uh, goes over to ChatGPT, you say something to it, and then it spits back an output for us. But with an AI agent, a similar kind of thing. However, imagine ChatGPT having predefined instructions on who it is, what it does, how it does that. And then we basically give it tools that allow it to actually do that. So if we wanna look at sort of an example of how an AI agent might look in this format, the input might be a new lead arriving via a contact form on your website, someone booking up a call, something like that. And then we're giving it 3 tools down here. So let's just say it's a LinkedIn research tool, a website research tool, and then an update CRM tool. So this is an agent that's gonna be able to, uh, research a new lead that come in. Find their LinkedIn, find their website, and then update CRM with that information. We would've told the AI agent that in the prompt. And then the output— that's meant to say output, says input for some reason— would be something like employee count, founded, location, obviously their LinkedIn information and some of their website information maybe. And then maybe even the agent just saying to you, hey, I've updated the CRM. So that's a really, really basic example of an AI agent. So then next on the agenda is why actually use these AI agents rather than just sort of doing it yourself or hiring someone to do it. So AI agent versus employee— I've done a little side-by-side versus here so you can see sort of the strengths and weaknesses of both of these options. So to win with AI agent, obviously they're cheaper in the long run, they're quicker, they don't get tired, unproductive, or bored, and they're quicker to train. And I just want to break down what each of those kind of means. So cheaper in the long run— when you build, let's say, this agent, uh, that we saw before, an agent like that, if you was going to hire someone to do that, that might cost you, I don't know, $1,000 to $2,000 a month. But if you was gonna pay someone like me or someone else to build an agent for you to do that, it might cost $3,000 to $5,000. So obviously upfront it's gonna be a little bit more expensive than hiring a person, but then in the long run, and by the long run I mean maybe 2 or 3 months, it's gonna work out cheaper than it would've if you'd actually hired that person. So that's what I mean by cheaper in the long run. Quicker, of course it can do, AI can do pretty much anything quicker than a human can. Uh, it doesn't get tired, unproductive, or bored, and also just works 24/7. So obviously with a human employee, this is just normal human nature. Sometimes you're gonna get sick, sometimes you're gonna get bored, you might be having a bad day, you could be unproductive. Obviously with an AI agent, it's a machine, so it doesn't do any of that kind of stuff. And then quicker to train. You can just dump a bunch of knowledge docs into it and it'll automatically know everything about you, all your processes, everything that you give it. It'll just instantly know everything about it. Doesn't have to remember it or actually learn it. Compared to an employee that would have to go on training courses, read your company documentation, and be there for a while before it sort of knows what you do, who you are, and how you do it. And if we have a look at the employees, one of the things it's better at compared to an AI agent, it's gonna be more flexible and easier to adapt. And these are pretty much the main things, uh, with a human employee compared to an AI agent is with that AI agent, like the one I just showed you, if you wanted it to just do something else, let's say you also wanted that agent to be able to email the good leads. So let's say it was scoring the leads and then all the good leads You wanted to send an email out to the sales team and let them know we got this lead through. Obviously at the moment it wouldn't have that functionality to email people, so you'd have to go in, make a tool, prompt it how to use that tool. With a human employee, if you just said, can you email this to someone? Even if they'd never done it before, it'd be pretty easy for them to figure out how to just email something over. So that's pretty much the main thing with employees is just you can say to them, hey, can you do this? If they've never done it before, they can kind of just learn how to do it pretty easily. And then another big thing is gonna be original ideas/human inputs. There's always gonna be things where humans need to step in. There needs to be human input, some human decision making. And obviously with an AI agent, it's not a human, so it can't do that. So that's the why for using AI agents. But then probably the most important thing is when do we actually use these AI agents? This is probably why you clicked on the video. You know, I want to use AI or AI agents somehow, but you don't know when and where to actually use these agents. So I developed something called the Routine Framework. Basically what this framework is, it's a really, really easy way for you to be able to identify where in your business you need these AI agents or where you would benefit from with these AI agents. Because I know from building AI agents for my own business and then also building AI agents for other people's businesses, that can be really, really tricky sometimes for yourself or for other people to actually know where in their business they can build these AI agents and where in their business is going to benefit the most from these AI agent systems. So I know you're all asking what does ROUTINE stand for? I'm gonna show you. So routine, the R is gonna be repetitive. I'm gonna break all of these down in a minute, by the way. So R is repetitive, the O is objective-based, the U is uncreative, T time-consuming, I input-heavy, N numeric, and then E error-prone. Some of these are pretty self-explanatory. You probably already know what I mean and where I'm going with these, but just to break it down a little bit further, this is it all laid out. So for repetitive, basically if there's sort of more than 3 identical steps thing that you're doing. So you have a process, let's say you're doing it every day, every week, or even maybe every hour for some things. If maybe halfway through it there's the exact same 3 steps, or at the beginning 3 steps, or at the end there's the same 3 steps or even more than 3 steps, um, then that's probably a good place for AI to come in and an AI agent to come in and automate that chunk or the entire process. And then O for objective-based. If there's a clear pass or fail, yes or no, or match criteria, this is also a good place for AI to come in because it means it doesn't have to do too much heavy lifting when it comes to decision making. If it's just like it says, yes or no, pass or fail, or matching a criteria, if it's just set in stone, things like that, that's a great place for an AI agent or AI system to come in and automate some parts of that process. And then U for uncreative, if there's no brainstorming strategy or original writing involved. So again, something like this is where an AI agent can come in and automate a good chunk of the process if it's completely uncreative, if it's just filling data into a spreadsheet or something like that where there's not really too much brainpower going on. You're not having discussions with other people, you're not bouncing ideas back and forth with team members or anything like that. If it's just kind of uncreative, boring work, that's a good place for AI to come in and see if it can automate some of that process. Again, T is time-consuming, and this is one of the biggest ones. AI needs to be solving actual problems. Everyone knows when you're running a business, time is money. If you can save time in any way, doesn't really have to just be of AI, but if you can save time anyway in a business, it is probably a good thing to do. So I've said 15 minutes on there. This really does depend where you're at with your business. It could be tasks that are really, really important and you have to do that takes longer than 15 minutes. Obviously still continue doing them if you feel like you need to. There could just be little things that you don't really need to do, but they're taking you 15, 20 minutes, maybe even half an hour. That is a great place for, um, AI to come in and solve these problems. And as you're watching this right now, you're probably already thinking of things like repetitive tasks that you do that takes sort of 15 minutes plus. And hopefully you're thinking right now, I don't need to do them anymore. I knew when I first started this, the first thing I thought of was all the tasks I was doing that were taking all this time that could be automated with AI. Those are the things I started with, and I recommend those are the areas you start with, sort of looking at the things that are basically time-consuming in your business. Everyone has time-consuming things, and starting there when it comes to AI agents. And then for I, we have input heavy. Again, I think I said spreadsheet earlier. Spreadsheet is a good example of this. Of just sort of inputting loads and loads of data into like a spreadsheet or a form or a field, anything like that. Anything when you're having to input loads of data into something is a good place for AI to come in. I've also found that when it comes to these large inputting tasks, a really good place for AI, maybe not completely taking you out of the loop or taking the human out of the loop with the inputting task, but just having them input way, way less and then having the agent take a look at those inputs they've already done and sort of be able to fill in the rest of the inputs. That's a really good place for AI agents, I found. And then numeric. I feel like this one's pretty self-explanatory. If there's anything that involves calculations, comparisons, pricing, reports, anything like that, I hate to break it to you guys, but an AI will probably be able to do that much, much quicker as long as it has all the knowledge on what it's actually doing. And it might take a little bit of training to get it right, but once it knows what it's doing, especially if it's doing sort of similar calculations or comparisons every time, it's going to be able to do that quicker than pretty much any human. So numeric, also very good place. And then error-prone is there's a high risk of human mistakes, things like typos or missteps. I know this happens a lot with, um, sort of contracts or proposals or legal documents where human error, there's not really any room for human error. Something gets messed up like a typo might look really, really bad or it might mess something up drastically. So having an AI come in and sort of do that, it doesn't really make mistakes when it comes to typos or anything like that. So that is a really, really good place for AI agents to come in as well. So now to help you out a little bit more, I've put together some examples and use cases so you'll actually be able to see sort of some real life examples of how AI agents are helping businesses and some examples of common roles that can be automated with AI agents. So this is just basically a long list of different roles and tasks that could be automated with AI and lots of people automating with AI agents at the moment. So I'm just gonna read them out to you. Lead qualification, auto responding to inbound leads, score them and route hot ones to sales. So This is a really, really good place for AI agents at the moment. A lot of people are jumping on this, these AI sales agents, these SDR, BDR agents, things like that. It's an absolutely perfect place for AI agents, especially if you're getting a lot of leads in and a lot of them maybe aren't so qualified, or your sales team is just having a hard time getting around to all of them, following up with them, things like that. Having an AI agent do it is absolutely game-changing. I've seen it firsthand in businesses. The customer support, this is probably one that we all know. This is pretty much everyone's starting point when it comes to, um, AI. Having an agent being able to automate all of your customer support is a really, really great thing to do, and it's really, really low barrier for entry when it comes to these AI agents. You can have the agent be a website chatbot, so the little widget in the corner of your website. You can have it via Slack, you can have it via email, anything like that. Pretty much all you need to do most of the time is give it a bunch of documents on your business or your company and sort of tell it what questions you usually get, how to answer them, everything like that. And then you're pretty much good to go with it. So customer support, really good one. Inbox management, again, this one's quite simple, but it's also quite handy for a lot of people, especially people that are getting loads and loads of emails every day and a lot of them are just rubbish or spam or ones that could be replied to really, really easily is a good one as well. Market research is also a fantastic use case for these AI agents because it could pretty much do this quicker than you and you're probably doing it anyway. So scraping competitive data, scraping trends, pricing, everything like that, you probably just do this on accident when you're maybe launching a new product. You look it up, see what other people are doing. When you're launching a new service, you probably do the same thing. So a lot of the time you just do this anyway. So having an AI agent to automate this all for you— I actually made a video not so long ago about a market research agent. So yeah, that is a great place to go as well. Invoice processing— I talked about this, uh, just a second ago. Having an AI agent automate things like invoices, also like contracts, legal documents, anything like that. That's a great place for AI agents. Internal help desk. This is something that's not used too much, but it's really, really good. And I've seen in businesses, this is a massive bottleneck of sort of employees or just internal people not knowing certain questions and having to send an email over to Dave over here, and then he doesn't know, so you have to send an email back over here, and then it goes back to Dave and back to her, and it takes a week at that point just to get an answer to a simple question. So just have an internal agent that has access to sort of company documents or whatever data you need, and then just having everyone be able to just chat to it and asking questions is really, really good. Data entry and cleanup. Again, this is kind of the same thing as invoice processing, just processing documents, contracts, legal documents, things like that. It's a great place for AI agents because it doesn't really make mistakes unless it's been prompted wrong or actually made wrong. Content repurposing. Again, this is good because a human is still in the loop here. So a human's still making, for example, a blog post or a YouTube video, something like that. And then an AI agent is just taking that content, repurposing it and then republishing it on say LinkedIn, Instagram, X, that kind of thing. Calendar coordination. This is kind of similar to inbox management, just like general admin stuff. It's really, really good for AI agents because it's something that everyone has to do, big or small business. Everyone at some point kind of gets bogged down with these admin tasks. So having an AI agent automating that is really, really handy. And then proposal drafting. I've already talked about this, kind of the same thing as the data entry, the invoice processing. That kind of thing. Having your AI agent just hand documents and data like this is really, really good. And then also put together a little process overview for you guys. If you were looking to get an AI agent built for you, an AI agent system built for your company, whether that's through me or through someone else, I just wanted to give you sort of a breakdown of how this process usually looks because this is such a new industry. Um, a lot of people don't really know how the process actually looks for getting these AI agents built for your business. So the process overview is pretty simple, to be honest. Um, everyone will do a discovery call with you to begin with to just sort of have a casual chat with you about the problems you're facing, what ideas you've had for AI agents if you've had any, uh, anything you've tried in the past before, that kind of thing. You just have sort of a 15-30 minute discussion with these people and talk to them about some of the problems you're facing. Like I said, you got any ideas for agents you've already thought of that you want, run through that, that kind of thing. And then afterwards it goes on to the process mapping, so mapping out how this would actually look for you. Sort of seeing if it's feasible, seeing if anything needs to be changed or worked around, which does sometimes happen. And then a prototype. So this is just going to be sort of the first version of this. And well, a prototype— prototype means a prototype. It's going to be the first version of this, and usually this will be sent over to you, and your agency, whether that's me or someone else, will say to you, you know, give this a try out, what do you think of it, if there's anything you want changed. And then obviously you go through that. If there's anything you want changed, then you can let them know and they'll change that for you. And then comes a fine-tuning. So there's always going to be some little things that need to be done before this actually gets sent out to production or is able to be deployed. And then after that is the deploying. So that's kind of an overview. If you didn't know, that's like a roadmap of how it would look if you decided today, I want to get an AI agent system built for my business. That is kind of how it would look. Of course, this might vary depending on who you talk to, what agency you go with, but this is pretty much how I do it, and this is pretty much how Uh, most agencies that I know do it as well. And then lastly, I just wanted to go over some common mistakes and how to avoid them because I've seen a lot of new people starting in this industry. I've also talked to a lot of businesses and company owners like you guys that might be watching this video, and I've seen a lot of the mistakes they make when either building AI agents for themselves or inquiring to get agents built for their companies. So the first one is trying to automate tasks that need human input. There's a lot of the time gonna be, uh, systems and there's gonna be tasks and roles that will need some type of human input, and platforms like end make it really, really easy for you to have human input. You can have, uh, whatever the agent does, you can have it send to a human and say, is this all right? And if they say yes, then it can carry on with the flow. If they say no, change it, they can take that feedback and change it. So trying to automate things that do need human input and kind of being scared of having human input in there and just wanting it to be all entirely automated by AI, sometimes that can be an issue when creating these systems. And the second one is not automating the basics first. So I've seen it when people will try to automate these sort of weird complex issues they have in their businesses without automating sort of the basic stuff first. So for instance, if you're really, really bogged down with admin work and just stuff that as a business owner, you shouldn't really be working on yourself, or you guys know the stuff as a business owner you shouldn't be working on, you should be delegating these things. And then they'll come to me and sort of want to automate these really elaborate, weird, complex, a lot of the time minor issues. And it's going to be like a really big project, quite expensive project. And they don't have any of the basics automated first. That's one of the things I see. And then this is one I see all the time. Number 3 is chasing hype instead of ROI. There's a lot of hype around the AI space and specifically things in the AI space. There's always hype around new things. So for example, not too long ago, there was a massive hype about things like MCP servers, things like that. There's also been hype around personal assistant agents, uh, customer support chatbot agents, things like this. Uh, the customer support chatbot agents are good, by the way. I dunno why I chucked that in there, but there are just things that you'll see pop up randomly everywhere and there'll be like massive of hype around it. Businesses will try and jump to implement these things into their businesses. But when you actually look at the ROI of these things, it's not usually that great. It's usually just a hype thing. It's usually just things that are like cool rather than actually useful. And I get it, AI stuff is cool. I still think all of it's cool. But as a business owner, you kind of got to look at the ROI of your investments. And there's a lot of things that are just sort of cool and just hype, and there's not actually much return on the investment there. And then number 4 I see is lack of knowledge leading to underutilization. And what I mean by this is a lot of the time companies come to me with sort of an idea of a system that they have in mind. And then when I actually look into the system, they'll have certain parts of it where they'll have human input in there when they really don't need human input in there. Or there'll be things that they think the AI agent can't do, so they make workarounds and they make compromises. But in fact, the AI agent can do these things. So I'm not asking you to, if you did come to me, be an absolute expert on AI. But sometimes when people come to me, they do have a bit of a lack of knowledge, which just leads to an underutilization of the AI Agent system, or potentially an underutilization of the Agent system. Because of course, when they come to me with these things, I'll go through the plan and I'll look and I'll say, hey, the AI Agent can actually do this. We don't have to compromise on this. We can do it the way you wanted it. So don't worry if you're coming to me with a lack of knowledge. I will help you out there. But if you are looking to build these things on your own, make sure you have a decent amount of knowledge so you're not underutilizing these tools and these agents. Okay guys, so that is it for this video. I really do hope you learned something. About building these AI agents and how and when to actually use them in your business. Because I know for a lot of business owners, it can be sort of overwhelming when you come into the AI space and you know that you want AI in your business somehow, but you're not really too sure how to implement it and also where to implement it in your business. So I hope this guide helps you. Of course, if you have any questions, uh, you're still not too sure after watching this video about if you need AI and where AI can go in your business, you can just book up a call, or you can just drop me an email and I'll help you out. Yeah, that is it for this video, guys. If you got any video suggestions you'd like to see, please do just drop them in the comments and I'll try my best to get around to doing them. I'll see you in the next one.
More from Rory
Dynamic Workflows are crazy. This is how to actually use them (beginner's guide)
