[(0:00) Blessing Richardson]: You didn't start your business to manage tools. You started it to serve people and deliver real results, but now the tech feels tangled, the systems feel scattered, and most days, you're the only one who knows how it all fits together. What if that changed? What if technology became your most valuable
[(0:15) Blessing Richardson]: tool? I'm Blessing Richardson, and I'm talking AI automation apps to help service based businesses leap into their tech enabled future. Earlier this year, I did an episode where I said that we're gonna see AI teams
[(0:30) Blessing Richardson]: and AI agents come into the fold probably next year. I think I was wrong. I think we're seeing it happen right now as I speak. It is time to start thinking about AI as a hired team member now. Well, I thought it would be later. People were asking
[(0:45) Blessing Richardson]: me for it now. So that's how I know it's time and these questions and these asks, and this work that I'm doing now is not coming from deep within the bellies of tech. It's coming from everything outside of tech actually. So when we start to think about team
[(1:00) Blessing Richardson]: members, the tech word for that is for so agents, right? And people are asking me to help them build that. If you've been sitting here thinking, I know how I'm supposed to use ai, but what do I actually do with it? Or you're like, I actually don't have a clue how I should be
[(1:15) Blessing Richardson]: using ai, but clearly everyone's using it. This episode might help you begin to wrap your mind around it. So a few things that have changed this year, really in the last couple weeks even, is that OpenAI, the people who make Chacha BT released something
[(1:31) Blessing Richardson]: into their ecosystem that allows engineers to build these agents, agents that can't observe work, they can decide what to do and they can act. So this is a bit more than writing prompts, still uses prompts, but we're building
[(1:46) Blessing Richardson]: AI in a way where it can see the work that needs to be done and then go do it. Google launched agent-to-agent, and they call that a two A for short, and it's the framework for having one agent talk to another. Now, up until
[(2:01) Blessing Richardson]: this point, you've probably heard me talk a lot about ChatGPT and GPT, so I'm gonna switch the paradigm a little bit. So it tracks, when we use chat GPT, we go in and we talk to it, we type a message, and then we wait, and then it responds back, and then
[(2:16) Blessing Richardson]: we read that message and we respond, and then it receives it, and then we wait and it responds back. And we can think of this as almost like a game where we're each taking a turn. Agents are different because it takes the turn itself. They begin to
[(2:31) Blessing Richardson]: talk to themselves and they don't need a human to wait on a human or even need messages from a human. But a human can kick off the conversation, say, Hey, I need you to go and write me a brief about these three things. And you can have one agent that's really, really good
[(2:47) Blessing Richardson]: at writing steppen and say, okay, I can write the brief, but I'm gonna ask my copywriter agent to help me write the right brief with the right context. And then the copywriter agent can say, Hey, I need to go now. Talk to the designer agent because the designer agent's
[(3:02) Blessing Richardson]: gonna be a little bit better at describing what kind of art we need for this brief. And so they get to orchestrate around themselves, take turns with themselves to keep working, and then they take a little bit longer to respond, but just like a team of people would, they're gonna take a little bit longer
[(3:17) Blessing Richardson]: to bounce around each other and get something done and get back to you. Now, I'm not gonna pretend like OpenAI and Googles are the first people to make this happen. As a matter of fact, there've been a lot of people, thousands of people working on this for many
[(3:33) Blessing Richardson]: years, and LLMs like Chat, GPT expedited this. But we're now at a point where this technology isn't just for engineers and developers and super super tech companies or companies with bajillion dollar budgets, right?
[(3:48) Blessing Richardson]: They're now building versions of this technology into new things like open AI's, agent SDK, and Google's A to a system, right? They're now building these new ways of using the tech that makes it even more likely
[(4:03) Blessing Richardson]: for non-technical businesses to get pretty buttons to click on to use the same tech. And that's a lot of the work that I do, right? When we build apps, we build products, it's taking all this power with tech and putting it behind. I like to call the pretty button, which trust me, it
[(4:18) Blessing Richardson]: can take months to make a button work. So I'm not minimizing the button. And actually in opposite, I'm saying that the button is a powerful expression of how simple it can be to leverage these amazing tools. And we're at the point now
[(4:33) Blessing Richardson]: where AI agents and automated workflows are truly being able to do work on your behalf at the pushes of buttons or even the utterances of words because of modern ai. So this is no longer the five year vision, it's no
[(4:48) Blessing Richardson]: longer, you know, the 10 year goal, or there's the sci-fi dream. This is as present as now. And so LLMs, we went from telling tools what to do to describing it with prompts, and now we're gonna let agents come up right after that and we're
[(5:04) Blessing Richardson]: gonna have them in a place where they can watch and now act. So what does all this have to do with me? Right? And you're probably wondering, asking the same question, what does it have to do with you? And people like you have been asking me about taking these apps and micro-tools, these internal
[(5:19) Blessing Richardson]: workflows and automations that, and these internal apps I've been building for their businesses and turning them into things that work more like agents. So this is where I am, and I agree
[(5:34) Blessing Richardson]: that this is a better way to build technology today because in truth, if I'm gonna build a bot that can write copy for me, then it's gonna have to have some skills. And up until this point, call it a bot or GPT or even an automation. And I know
[(5:49) Blessing Richardson]: these things are technically all different, but from my perspective as a business owner, when I'm not an engineer, I don't really care what it's called. The whole point is that I'm gonna tell you to go do a thing and I want you to go do it. So depending on whether or not it's an automation or a GPT or a bot,
[(6:04) Blessing Richardson]: that's all minutia that at that moment in time, I don't care. I just want it to work. So when I use 'em interchangeably, that's why. Okay? And so if I have an automation on the other side of this, somehow I'm going to give a description of some piece of copy. I want it to write, it needs
[(6:19) Blessing Richardson]: to know like how do I need copy written? Now, AI's really good at just writing things, but specifically for my business, if I want you to write a blog about this very podcast episode, what does it mean to write a good blog for Sayla my company? Well,
[(6:35) Blessing Richardson]: I have to tell it that I have to give it those requirements, okay? And so we know that clearly I can go to ChatGPT and say, Hey, here's my podcast episode, write a blog. But like what sections do you need? Are there tools and apps that I referenced in the podcast episode that probably need
[(6:50) Blessing Richardson]: links for people to go to? How do I give it instructions on where to find that information? And I can tell it where to go, but can it actually go there? I can tell it, Hey, go to my Google Drive and go find my spreadsheet of all my links and then pick the best ones for the things I said in
[(7:05) Blessing Richardson]: the podcast. But then how does it actually go to Google Drive? Well then we have to create tools for our agents, right? We have to give these bots or these agents the tools to actually go and do things, not just say things back at us and then we
[(7:20) Blessing Richardson]: have to plug it into our systems. Because chances are you can't give it that much data all at once because one, it's a technical limitation to some degree. And two, if you give it a lot of information, you have to find different ways of doing it. So it's really easy to say, yeah, I'm not just gonna give you a pile
[(7:35) Blessing Richardson]: stacked high of all my different transcripts and all my possible different links and instructions on how to go to Google Drive. Like no, I'm just gonna tell you like use this link to go to Google Drive and go get the file. So we have to give it a tool to be able to do that, right? And then
[(7:50) Blessing Richardson]: when it goes and it writes that blog for me, how do I know it's done? How do I know it did it the way that I wanted it to do it? I would probably want it to tell me like, oh, what a person. Funny enough, as a matter of fact, how would I give this to that person? I would say, I just finished recording this
[(8:05) Blessing Richardson]: episode. The episode is in Google Drive here as an audio file. Please go take a listen to that episode, turn that into a blog post following this template that I've given you. And by the way, go to this spreadsheet to get all the links that you might need to write some
[(8:21) Blessing Richardson]: decent show notes for the episode. Can you let me know when you're done and upload that? The blog post and the show notes back into the same folder with the audio file. That's probably how I would describe that. So if I'm gonna delegate that process to a bot
[(8:36) Blessing Richardson]: or an agent, I need to one, give it those instructions. Two, give it those templates. Three, give it access to the information it needs. Four, connect it to all those things to do it. And so this is how we start to think of our I team members, right? Essentially, if I give you all the same inputs
[(8:51) Blessing Richardson]: for a process, I would a person then can you produce similar outputs that a person would produce? And so then that brings the question, maybe we should be looking to hire AI and not just simply find little automations to tell it what to
[(9:06) Blessing Richardson]: do. If I look at AI like a hire in my business, then I own the responsibility of onboarding it as well. I've given it all the tools and information it needs to do its work effectively. If you haven't heard it in that way, I hope
[(9:21) Blessing Richardson]: it starts to jog some ideas and inspire you to think differently. If you are at a point where you're like, okay, this is a lot, it's AI everywhere. I've noticed that AI showing up in my slack, in my Google, in my notion, everywhere I go, there's an AI feature, I can't even turn it off. Well, that actually is kind of frustrating.
[(9:36) Blessing Richardson]: There are tools that I don't even wanna use AI in, and I can't seem to turn it off. So I get that. But there's a reason why it's so pervasive. It's showing up in a way that makes me almost feel like it's kind of like electricity. Sure, you can live in today without
[(9:51) Blessing Richardson]: electricity. There are still people who do to some degree, and even then we might argue that they still need a little bit to survive with the outside world around them. But you can still live in today's world without electricity. It's possible, but it definitely changes your experience. It changes
[(10:07) Blessing Richardson]: your whole existence when you're trying to go out to meet people. And 99% of the time, everyone that you're going to meet is using a vehicle to get there. They're using some form of car that's running on some sort of fuel to make their
[(10:22) Blessing Richardson]: transit times maybe more like 30 minutes, 45 in traffic instead of like four hours, right? So it's not that you can't live in this world without electricity. It's just very different and sometimes really hard. And I think that's where we
[(10:37) Blessing Richardson]: are with ai. We're moving to this place where you can live in this world without ai, but it's becoming so pervasive and the more understood it is, the more used it is, the more it's becoming fundamental to how we're existing. Well, I do think we should be judicious about
[(10:52) Blessing Richardson]: how we adopt it and where we do. I believe we're beyond the point of deciding whether or not we should use it if we want to exist in the modern world. I think the answer is yes, we should. But how and what is appropriate? When I think of team members, it's the same thing. What type
[(11:08) Blessing Richardson]: of tasks and roles are appropriate for an AI to do within your business, on your behalf, and with your team? Begin to ask yourself those questions because you're gonna start hitting these walls with AI being everywhere. And so having an idea of what you're gonna opt and not out of
[(11:23) Blessing Richardson]: before you get there makes it really easy. And also saves your team or your business from falling into operational tailspin. So just begin to ask these questions, whether or not it's just to yourself or to your team members, or if you wanna have some fun, maybe have the conversation with Chad GBT. I wouldn't. It
[(11:38) Blessing Richardson]: might be a little biased. So to bring it all home, you know, for me, this isn't a pivot. It is just finding more language to better define the types of solutions I'm helping people deploy in their businesses. So you're gonna still hear me say things like micro-tools and
[(11:53) Blessing Richardson]: apps and data and automation and AI and all these other things, and bots and agents, and get more clear around that. But you might hear these things under the umbrella of a teammate that has a role, that is a
[(12:08) Blessing Richardson]: bot, that is an agent, right? You know, start to see that umbrella of the I team member a bit more. 'cause I think it's a good umbrella and it works really well to explain how we are shifting to use AI in our businesses. If you wanna see what this looks like, I invite
[(12:23) Blessing Richardson]: you to join me at SIL two slash go live where you'll be able to sign up for some of my webinars that we're gonna be doing about using AI like a teammate, like a team member that works with you and alongside you. So
[(12:39) Blessing Richardson]: yes, we're gonna get into the fun stuff about how to create tools that your agents can use to go do things within your systems with your business data on your behalf. But we're also gonna spend more time focusing on how do you create the data itself? How do we store this information
[(12:55) Blessing Richardson]: so that the agents can actually have access to this to do things the right way? How do we word the instruction so that we're being effective not just in the information and data, but we're being effective in instructing it to do the things that we need it to do. You are gonna get a lot more around that
[(13:10) Blessing Richardson]: and so that way you can work into your own business. The understanding that as we do things and as we come up with processes and rules and systems in our business, that we now have this AI teammate that along for
[(13:25) Blessing Richardson]: the ride, and we need to be able to give it the right data and information so it can continue to do its job.