I LOVE AI Month With Noelle

spotlighting myself. So if you are not speaking, you won't be on camera. But if you do ask a question, provide a comment or feedback, you know what, there's like a— there was a show when I was a kid. It's like you're on Candid Camera. So you'll be on camera. And so I just want you to know that. So that is the good news. Welcome everyone. This is our weekly get-together that allows us, those of us who love AI and are using AI, as well as those of us who are on a learning path to understanding what AI can do for ourselves, for our business, for our families, for, um, our companies. This is a great opportunity for us to come together, get to know other people on the journey so we don't feel quite so alone, which sometimes we can feel, but also to give each other ideas and, uh, get a chance to, you know, maybe even offer our help or support to those who need it. Or better yet, ask for help and get support when we need it. So this is a great, uh, free community. It's called I Love AI. You can easily come join us. This is a Zoom call, so on this call we get to interact and have some fun. We're gonna break out into small groups and get to know each other individually, talk about a topic, or just share what's going on in our worlds. Um, and if you wanna be part of that fun, you do have to be here in the Zoom, and that means you need to be part of the community. But it's easy to do. You just have to go out to noelle.ai and that's it. My name, N-O-E-L-L-E dot AI, and you are there. Just click the link. To the free community and you're in. And we just completed today our, our I Love AI Challenge. This is a challenge we have run. This is our third year. And it's really, really exciting because this year we focused on building AI employees. Now I use that word carefully because sometimes when I say employees, people are like, wait, aren't humans employees? And are you getting rid of humans? And No, not at all. Um, what I want is just, I wanna move away from AI as a tool, AI as like, uh, what I call vending machine AI, where I ask AI for something and it gives me something back and it's very transactional to this more thoughtful approach of what if I designed a human AI experience for my company where the humans are intentionally doing certain things and AI is intentionally doing certain things and we all live happily together. Together. Symbiosis, right? So that's what the premise of this challenge was. It was 5 days. We had over 500 people join the challenge. Um, we ended up creating a VIP section, um, uh, for the challenge. And in that VIP section, I wanted to just deliver as much value as I could. So I took all of my— I mean, I've been using ChatGPT for like since I think it said 1,500 and 87 days or something like that. Maybe it's 1,600 days now. Um, and I've been using a lot of these technologies for longer than most. And so I have fine-tuned my models to behave a certain way. And so I am sharing exactly what that is. So for those who sign up for VIP, they not only get, of course, the opportunity to just ask questions in a very close-knit group, but we also, uh, you get the benefit of me just sharing with you all of my prompt, uh, repos. All of my examples, all of the, the things I've learned along the way, the documentation, um, and most importantly, how to like actually deliver AI projects successfully. So that was the premise, right? Is that originally, I mean, we now know that most projects fail when you start building an AI project, most of them fail. And the reason why is a thing I say all the time, right? Part of my ethos, the reason, like you could see on my book, there's a tiger on my book. It's about baby tigers. It's that when we start building AI, it's very cute and cuddly and it's small and useful, but it often, it, it evokes this kind of really excited, like, I can't believe this is happening. This AI is so cool. I can't wait to show people, show my mom, post it on Facebook. This is amazing. And the reality though is that that little system is going to start growing up, getting more use cases, being asked to do more things. And the smaller the model is, the more appropriate time it is to ask really important questions. Um, and that's why I use a baby tiger kind of as an example, cuz I'm like, when you see, uh, some of you have seen like a dog, when you see a dog, uh, and maybe a puppy and someone's like, oh my gosh, look at my cute little puppy. The first thing I do as a dog owner is I look at the paws and I'm like, wow, those paws are huge. Do you know what you've gotten yourself into? Like, I don't say it to that person in case they don't know and I don't wanna be the one to tell 'em, but That's the first thing I do. And that's what I encourage us to think about with artificial intelligence is when you adopt a baby AI model, when you go and use a model that you don't know very well or you haven't evaluated, like look at the pause, right? See how, how big is this thing gonna be? How much is it gonna cost me? Like do some thoughtful foresight. Uh, and that's what this week was all about is like, how do we think about something when it's small? And in your mind you're gonna be like, I, This is overkill. I don't need to do this level of examination right now. But AI is one of the first technologies that can massively scale with very little effort. And that scale is what scares people, right? That's what we're talking about with uncontrolled AGI. Um, so I really think, you know, this, this metaphor has served me so well because I don't only ask, you know, how big its paws are and how big it's gonna be. I also, Ask about what it's gonna eat data, right? How much it's gonna eat, where is it, where is it gonna come from? Am I paying the people that are giving me that data? Like all of these are really important questions when it's little, when you can mitigate the risk, when you can solve those problems, when you can plan. It gets very hard when that tiger grows up and becomes, um, you know, a, a unwieldy system, which is what we see. I have two Google Alerts that I always encourage people to at least have two. I have lots, but that's two. Of them is AI for Good and AI Gone Wrong. And I like to, on a daily basis, get visibility to how people are using AI, but I wanna know the best parts of AI and how it's saving lives and protecting people and supporting justice. Like there are, there is the, the one misconception I think we, we sometimes have, or at least people that I talk to have, is, is that it's like one big AI in the sky and that AI is controlling all the things and we should be very worried when actually the reality is It's millions of AI and these millions inside that million are such genuine, important, impactful use cases that will bring tears to your eyes that they're so, so useful, so helpful, um, that they're changing people's lives in the most positive way. But then the alternate is also true. So I wanna know that too. You, some of y'all know me, I'm optimistic by nature. I tend to be looking at the bright side of life. In a Monty Python way, right? Like I'm always kind of like looking at things glass half full, but I am a rational optimist, which means I do like to ground myself in the truth of what's happening in the world and stay up to date. Not too up to date, cuz I don't wanna get like, you know, anyway, but I like to know what's going on. So I have this alert on AI gone wrong and how AI is, you know, when people take a baby tiger and throw it throw it in front of their customers. What's the worst thing that could happen? Oh, you could sell your truck for a dollar, or oh, you could accidentally manufacture refund policies to your customers. Like, there are bad things that happen. Worse, you could lead someone to make a very bad life decision through an AI model you just didn't realize could do that to someone. There's very drastic dark sides of what this technology technology can do, but I'm a firm believer that we're the, we are in control of that right now. It might not be true 5 years from now or even 5 months from now, but as it states right now, today, the best control you have over a system is you actually getting involved in helping it learn, right? I said in my TED Talk, I encourage you to go listen to it. We just passed like 600 or 700,000 views. So go check out my TED Talk. But in that TED Talk, I talk about us being the, the first generation really to be able to go in and build AI systems and teach it what good looks like and teach it like how to solve note-taking and how to solve, you know, financial models. Like we're this generation of people that are using AI in a way that's never been used before. But I pontificate, I hypothesize that we might be the last generation to actually teach these models about thoughtfulness, empathy, about You know, pause, intentionality, um, about making sure that we are prioritizing trust, security, safety. I don't know how many of you know this. I talked about this on our call today, uh, with, uh, in the challenge, but I grew up on the golden age of science fiction, right? I grew up there. Uh, and so I read hundreds of science fiction books that were written between 1930 and 1980. Maybe, and a little '90s were thrown in as I got older. Orson Scott Card, for example. Love those books. And like all things, I love the books. I'll just keep it at that. But anyway, I'm reading these books. And of course, Asimov creates this concept fictionally before we have AI running amok, before we have little baby tigers scattering around. He says like, hey, when you build technology, you might want to, I don't know, create the three laws of robotics. You might want to create some systematic process that tells every robot, every AI system, every agent how to be safe, secure, responsible. And so that is the way that we started our challenge this week is that, and actually I think I have a deck, so I'll show it to you. But the way we started this challenge, which it was called How to Build Your First AI Employee, but I'll share with you the reason that we did that was because the first step when you build an AI system, the first thing you need to do is figure out why you're doing that. What is the purpose? What is its job? Why should it do what it does? And when I say that, I kind of equate it to being an employee, right? If I'm gonna go hire, right, the first thing I need to do is have a job description. And I don't think AI should be any different. The jobs I have AI do will be different than the jobs I have my humans do. It will be intentional and designed human AI experience. But eventually I do want to disclose and document exactly what my AI system is responsible for. Now, fun fact, I find the documentation of job descriptions for humans is not all that great. I find many times that people have copied and pasted job descriptions and some of this stuff doesn't even make sense. How many of you have seen a job description asking for 6 years of GenAI experience, right? Or, you know, like, You're like, um, it's not Java, right? It's generative AI. So I feel like there's a bit of lackluster support for really focusing on, if I'm gonna hire a human, what work are they gonna do and how are they gonna do it? And we should have that same level of meticulous focus when we build an AI system. So that's what we focused on on Monday and Tuesday. We drafted the job description of our AI employee. What will it do? What do I as the human wanna keep and what do I want this AI system to take? And it is 100% my choice, right? The AI system's not gonna come to me and be like, I want all your work, I wanna take over your job, I wanna put you outta business. Like it's sitting there waiting for me to tell it what to do. So it is in that, and I, uh, mentioned during the challenge, I use a process called Socratic prompting where I don't use vending machine kind of AI where I'm like, build me a social media post or generate an action item list. I will instead start with the problem. I would like to engage people who love AI like I do. I wanna engage responsible AI leaders in my community, in my industry. I would ask the model, how should I do that? What are the top ways that I should get in front of these people? And I hope it says social media, but I'm not going to like feed, you know, what do they— I was about to say feed the witness, but that's not the saying. I'm not gonna lead the witness, right? I'm not gonna lead the AI to the outcome because maybe I want the AI to think differently than me. Isn't that the idea? Like, can I get diversity of thought? Um, I have a way of thinking. I have people around me who provide me diverse perspectives. I also have an AI system, if I ask it the right way, who can help me see the world in a different way, in a maybe more positive way, a more ethical way, in whatever way I actually choose. And so that first step is not what model will I use, what cool tool, oh my gosh, Opus has a million tokens, what can I do with that? It starts first with what am I trying to achieve? What problem am I trying to solve? And how will I measure the success of that problem? So Monday and Tuesday we did that, and these are all recorded. The recordings are phenomenal, VIP, like Oh my gosh, crazy. We're synthesizing this into a book. It's so good. Um, but if you want to take it, you could still go back and watch the replay from this week. It'll be available indefinitely, so you can always go back and do it. But these— the Monday and Tuesday was really about how do I get my idea of what I want, my way of measuring success on how to achieve it, down on paper or a notepad or in a markdown file. So that I've documented what I want before I've ever built it. Then, this is Noelle, I do this, you don't have to, it's not even the right way, right? But Noelle's way of doing it is I take that information, I'm an engineer, I'm also a technical program manager, so I truly appreciate the art of, wait for it, business requirements and functional specifications. Like, this brings me joy when I see them. Do I need them? Maybe not, right? Maybe AI can extrapolate all that,. But it gives me a clear chain of thought of how I'm building what I'm building and the outcome I'm expecting. And I'm able to now more clearly articulate that to the model that I'm using, the AI system that I'm using. So that's what we do, right? We create the job description, we define what I call the, the system prompt or the meta prompt that will embody all of those business requirements and functional requirements. And then I build it. And so I like building. I would prefer to build a lot of things in a little period of time than one thing or nothing in a long period of time. So we ended up, I uncovered a platform, uh, and I ended up using this platform called Marbleism. Um, and Marbleism is a platform that gives you 6, like de facto standard kind of business agents, right? Um, and maybe this is a good time. I'll go in and share my screen so you can kind of see what they are instead of just hearing me. Talk about it. Uh, hopefully, uh, there we go. I'll put it in slideshow mode, maybe. Oh goodness, I hope you can see it. All right, there we go. Um, so this gives me the ability, right? And I'll, I'll go through— I just talked about the first few things, but I thought it would be cool for you to kind of think about it through the problem lens that our challengers were going through, which is— I'm— this is Noelle Russell Russell. I'm in an, uh, business owner. I have 3 companies. I have tons of email, more than 100 plus, way more. I don't even want to say it, um, right? We actually had a little— we were gonna have a little competition of who had the most unread emails, and then I was like, you know what, never mind, I'm, I'm getting nauseous thinking about it. So we decided not to do that. But if you're one of those people who's like, oh goodness, zero inbox is a dream I don't know how to achieve, this was, this was a problem scope, right? Like, hey, wow, what if I could solve that problem? What if I could post consistently on social media? And I mean every single day, twice a day on 5 different platforms, different content with value and ideas that are in my head. I mean, if it was just, if I could just think it and make it happen, I would have so much rich content out there. But instead I think it and then go build it and it takes me a long time. So that is an opportunity. Blog. Some of you are in my AI Leadership Journal, uh, which is my 17,000-member newsletter on LinkedIn. Love it. I love it. I love writing. I like writing long form. Obviously I wrote a book, like, I like it, but it takes a long time. And so I was like, this is another problem. I'm not consistent. I try to do it every week, but it usually turns into once a month. And then in December, some of y'all know, I write one every day for 12 days to make up for it. It's called the 12 Days of AI Leadership. And I'm like, every single day, um, which is good or bad, but I'd rather have consistency, right? Not ebb and flow with like what the volume of my business, but instead just consistently show up for people who sign up for my newsletter and wanna hear what I'm talking about and hear what I'm learning. There's mechanisms, AI systems that help me do this. Reaching out to customers, right? Usually as a business owner, me, I'm the CEO for sure, but I'm also the marketer, the copywriter, the salesperson, the accountant, right? Customer service rep. Um, you know, uh, the SWAT team. Um, IT. I'm all the things, right? I'm the CISO, the CEO, the CRO, the, you name the C and the O, that is, I do all of that. And I do leverage some humans to help me, but I have a lot going on as a small business owner. So that's what what the premise was for the challenge, and we wanted to look at an opportunity to fix it. So the first thing, as I mentioned, we did problem definition. So we spent some time, you can go watch the video if you join the community. Day 1 of the challenge is how do you define the problem that you wanna solve with this AI system? An AI employee is not a like super AI that does all the things. It's narrowly focused on a job, a task, a specific thing you want it to do. In my case, an executive assistant that manages my email and calendar, or a social media bot, right? Agent that publishes short-form content or a long-form content, or a legal assistant, right? Like these are all different parts of my business that I currently struggle to val— get value from. So this is what we ended up doing. We went from describe the business to refine that into business and functional specs. Now, if you know a little bit about functional specs, which I get it if you don't, because it's a dying art. Um, but it's when you go through and you think about everything you need to think about prior to building. So you're thinking about security, privacy, trust. You're thinking about where your data's gonna come from. You're thinking about all the baby tiger questions. How big will this be? What's the infrastructure gonna be? How do you make— who's the audience? How are they gonna use it? How much are they gonna use it? Tokenization. Like, these are like the big outline of how you're going to implement before you've even selected a model. So this is what we did on day 2. Then we got to day 3 and I was like, okay, let's have some fun. So I started talking about tools on day 3 of the challenge. If you go back and listen to it, you'll see I talked about a bunch of tools. I mean, we went through Claude Code and Lovable, and I mean, you name— okay, I won't say that, but you name 8 and you might hit a few. Um, but Crew AI, uh, Poe.com, like we just went through a bunch of different tools that could help you deliver on the promise of your functional specifications and your business requirements. This, I think, is what's missing, right? We saw a study recently that said that, people that are building AI systems, that 95% of AI pilots are failing to re— they're not failing, meaning they're not breaking. They're just failing to return the value the business is expecting. They're failing to return an ROI or an ROE. T, right? Either a return on the investment from a financial perspective, or at least a return on time, right? A time savings. And usually the reason that is, is because we didn't think about that metric when we started building it. We kind of were like baby tiger mode. Yay, I— this tech does this cool thing. This is amazing. How do I use this in my business? I'm not really thinking like, how am I gonna measure that that's successful? How am I gonna know when I'm done? Definition of done. These are the things that shift your being part of the 95% that are failing and part of the 5% that are winning, is that you build it like any software, right? You start with the problem, you define the specification for how it should be built, and then you select the right model for that problem. Sounds like, uh, some of us, like you all, some of you who are listening, you're in my world, so you're already AI people., but many aren't. Like a lot of my clients that I talk to, they're not deep into AI technology, so they might be pretty loose in their business requirement and requirements and functional specifications on their current software that's not AI. And so I rekindle their love, uh, or hate of this process, but it helps you build AI systems that do what you tell them consistently over time, meaning that Baby Tiger is Reined in, wrangled, put a moat around it. I like moats better because I don't wanna put fence around it, but I'll put a little moat so that, you know, you don't hurt yourself with this technology. And that's what these specifications allow us to do. Now, AI employees, the word I use might be triggering to some because it's attached to like human definition. Many of us define ourselves by our work. I certainly used to. But the reality is, is that we are in a season with this technology of redefining what work looks like, redefining what AI does and what humans do, and that it's our choice to do it, right? Again, we're not talking about AGI, though you can talk about it. It's a little forward thinking, but right now, today, 2026, how are you going to do this? This is my hope for you. So I'm gonna walk you through what I did with the team and show you like what marbleism does. Now, for those of you in the chat, Uh, you probably saw it. I will share, uh, in the— you can go look, look it up, but I have a discount code if you want to play around with this, uh, that I'll put in the comments on LinkedIn and on YouTube, where if you use it, you'll get either 10% off the lifetime of your, you know, if you decide to keep it, but more importantly, 70% off just a month of use, which makes it about— someone did the math for me— $13 to test out. And I would encourage you to do it, even if you never use it again. Just play around with an agentic platform. There'll be many of these that come and go. I encourage you to get your muscle memory developing on how are you gonna use your words to control a machine. And this is a great way. I did a lightning session yesterday where I went through and trained all the models I'm about to review with you in about 40 minutes. I did 6 agents. It wasn't pretty, but it, we did it. Um, and I think that's the kind, I like putting my fingers on a keyboard and building stuff. That actually delivers some outcome, some value, and then I can go back and make it better. But minimum remarkable product is the way I think about it. I wanna build something small, minimum, like these agents I'm about to show you, but it needs to deliver something. It needs to be a remarkable thing. When I see it work, I want the emotion of like, oh my goodness, this is amazing. I can't wait to tell my friends, right? That's the emotion I wanna generate when I build an AI system for a company. At least in the first few days or weeks with them. So the first one, uh, this is, uh, the Executive Assistant. This was the first one I built on Marbleism and it like changed a lot about my ability to react, uh, to emails coming in. Um, you could see it drafts emails for you. You don't have to use them. It doesn't have the ability to send them without you. Responsible AI in action, right? But it requires a human in the loop, which means that you are required to be part of this specific agent's behavior. You cannot, it cannot do things on its own. It cannot delete email, for example. It can only archive email. However, what it does very well is it goes into your inbox and it triages them. And I'll, uh, maybe I'll do a, um, a demo, uh, in just a few seconds, but it triages your email, it labels your email, it identifies things that are urgent and not urgent. Some of you know, Over the years, I've missed some really important emails. Matter of fact, I identified a single use case cost me about $210,000 last year, a single use case of missing a certain type of email in my inbox. Needless to say, I became hyper-focused on solving this problem for myself, and so this is what I currently use to solve it. Um, that is one part, right? It also manages my calendar. It's an overall executive assistant. It also has the ability to communicate to all the other agents. I haven't really pushed the edges of what it can do. I'm just kind of following my personal pain and building out a system that matches that. Now, I will tell you though, um, I wonder if I have it. Mm, I don't know if I can show you. Maybe I could stop and I don't, yeah, I have a Word doc basically. It's 6 pages, 7 pages of instructions that I give to this model. So Marbleism is cool. It's another platform. There are lots of them. I'm not, picking one over the other. But I do think it's important to get, get into the habit of taking an agent and feeding it the right system prompt for you to control its behavior and realize just how much control you have through the words that you use. Every model is like this, right? If you go to Meta AI Studio and you build a little AI there, it's gonna ask you the same thing. What is its purpose? What is its description? How, what are the instructions? So this work of going in and providing instructions that control agent behavior, this is gonna become commonplace if you're not already doing it. So get into the habit, even if you just do this over the weekend and just build a bunch. And like I said, if I can do it, if I can give you a code that allows you to do it for $13, like some of you are gonna burn $13 on Lovable credits this weekend. So, so I think it's really, uh, a really economical way to gain some muscle memory experience., controlling the behavior of an agent. Now, this was just one of them. Another one that I really liked, um, is, uh, is the social media. So this allows you to feed it your content, your brand voice, your, uh, the information, your audience, like all the information that it needs, and it can, in your own tone and style, help you design content, uh, whether that's a content calendar, whether that's a campaign. Um, I used it to help me ideate on my I Love AI Month campaign. I wanted consistency and I wanted an evolution. I could ask it to do things that were very tough for me to articulate to like a human team, right? I would ask them and they'd come back and I'd be like, well, that's not really what I meant. But instead, and again, I, there are humans that do this and if you find one, grab on tight. But in my experience, I was struggling with that conversation. So instead I get this agent, I point it to my already existing social media and I say, That's what I want it to look like. These are the topics I talk about. Help me come up with a campaign that speaks to this audience. And it just allowed me to do what I was already doing better, faster. Drafts, scheduling, right? You'll notice here it, it actually can't post on my behalf, nor do I want it to. Human in the loop. When you start using agents, I always encourage you to be in the loop, meaning, and there are 3 different stages, right, for human AI autonomy, right? The first stage is human in the loop. Anytime an agent does something, don't do the— don't send it, don't, uh, delete it, right? I wanna be the one that ultimately makes the decision. Some use cases will always have a human in the loop, some of them. The next stage though is human on the loop, meaning that I am watching the loop, the workflow, which is the loop we're talking about. I'm watching the workflow, but I'm not actually halting the workflow. It's gonna continue to go and I'm just monitoring it. Maybe I even create, um, a con— constitutional AI agent. Maybe I have a red teaming agent that's like monitoring and checking how things are going, but I'm not stopping the process. I'm letting it go, but I'm observing the entire process. And then some of your favorite, um, a human out of the loop, completely autonomous AI. I'll just leave it at that. There's that. I like to have a little bit— I, I actually see it as a, a point of evolution. There are use cases that you're gonna— I mean, spam, for example. There are a few of us on the planet statistically that check the spam folder consistently to see if spam is doing a good job. But most of us just kind of say, if you got caught in spam— I mean, me, like sometimes I email some of you and my email gets caught in your spam. You don't go and look there first, right? Like you're just like, you didn't send me an email. I'm like, It's in spam. Go check it. Make me your friend. Add me as a contact, please. It's my personal email. Like, add me. But that robot, that AI agent, if you will, that's a very, very simplistic algorithm, is running and tracking spam. And most of us don't even care. Like, we don't monitor it. We actually just— we actually think we wish it would do a better job, right? Sometimes we get content that is spamming it anyway. But that perspective is helpful to think there will eventually become a process that you trust so much, you've watched happen on the loop so often that you're like, all right. And I always say, oh, I was started saying this earlier today, it takes a long time. Like at OpenAI, it was like 7 years, right? Before they were able to actually publish their model to the world. And, and even then it was a little bit broken, right? It takes a long time for models to show trusted maturity. And so you wanna give it that time and you can, again, you can get out of stopping the decision by being on the loop but observing it. I will tell you at Amazon Alexa, we created an accuracy review team and our job was to sample hun— I think it was over a billion utterances were coming into Alexa at the time. We would sample 100,000 and just spot check like that was human, I mean, we were like barely in on the loop, but we were still involved. We were manually looking, spot checking to make sure that utterances that came in were getting answered the right way. We weren't looking at, we weren't in the loop, we weren't looking at a billion utterances and seeing if they were correct, but we were spot checking. So we're still observing and checking, um, but not fully autonomous. And that was Alexa, right? Fully baked production system. Continuously had humans observing, monitoring, and checking. So it's kind of up to you which use cases you use. I mean, y'all know autonomous vehicles. Um, there's not a whole lot of, uh, I mean, I would imagine there's some spot checking maybe. I don't know. Waymo, like there's some interesting data at Waymo, um, that's happening. So these are the 3 levels and you have to choose where you're gonna be and you should know where your AI system is at all times. You should know, are you, do you have a human in the loop, on the loop, or out of the loop? Anytime you look at an AI system, it should be instantly identifiable what state of human interaction it's at. So that's what I, uh, like about these models is that they're walk before you run. This is not an autonomous, uh, CRM system, but it does provide you a little bit of a taste of what that might look like. Um, so I use this. I'll, I'll show you an example of it, um, in just a second. Uh, Penny is a long-form writer, so it's a little different. Deeper research, more analysis, making sure that your content represents, right, sounds like you. So it's asking questions and reiterating perspectives that you have identified. And again, you're giving it instructions. So these are all like little agents that if you train them well, they will start doing things that you're probably either not doing because you don't have time or doing a lot less of because you don't have time. And again, the perspective is through a business owner. We then have Stan, probably my favorite. Stan's the salesperson, sales representative, not a person at all, an it, not a he. Um, and so Stan is going to use publicly sourced, like Apollo, right? Publicly sourced databases with, uh, people who have consented and opted in for their information to be seen, which is what you do on LinkedIn, by the way. Um, so it uses LinkedIn as one of those. 700 million leads are accessible to Stan, and it's gonna go and look at your ICP. Your, your, uh, ideal customer profile. Now, it doesn't know what that is until you tell it. So you have to tell it. Now, what it does know is you're gonna point it to your website and it will extrapolate. If you do not provide any additional information, it's gonna go to your website and go, it looks like these are the people you're trying to serve. What I like about it is it goes out, it connects with them, and then it also starts an email sequence. Most business owners, small business especially, It takes them forever to build out an email sequence to just get people to raise their hand to work with them. I mean, there are already people who want to work with the— your business. They already wanna work with you. They just don't know you're there. They have no opportunity to raise their hand and get you to call them. And so this is a great, great way to do something that every business owner should do, but again, likely doesn't have time, um, or doesn't know how to do well. And that's a great kind of symbiosis between AI and humans. And then the last one is a receptionist. Again, I've done this the hard way. I have provisioned Twilio accounts. I have provisioned, you know, created, uh, voice AI agents from scratch. I was shocked when I opened up Marbleism and I onboarded Rachel, the AI, uh, receptionist, and it instantly gave me a, uh, a business line in the area code that I work in. Instantly, it now can book calendar appointments. It can have conversations. People can call in. It can answer questions based on what's on my website and drive them to book a call with me. Super valuable. It can also, in the middle of the call, be like, oops— remember, human on the loop versus in the loop— it's not even all or nothing. This agent has specific conditions that say if the condition is X, Y, or Z, transfer it to Noelle Russell's cell phone right now. And that's like, I don't know, let's say Netflix calls me and says, we're going to do a show about you. I don't want to miss that call. My AI agent better not be like, Netflix, may I please have your email address and maybe I'll call you back later. Like, no, I'm like, you get that to me right now, right? I want that E! True Hollywood Story. It's happening. So that is the differentiation, right? And that's why I like testing these things out. I go in, I test it out and I'm like, I know what it takes to build this from scratch. The fact that I can click a button, get an email, like get a, a phone number, and then that phone number can book calls for me and transfer calls to my phone. Like that wasn't easy to set up even a year ago. So really, really cool stuff. So that's what, um, I ended up doing in the challenge was demonstrating this. And of course, let's go get, get fun. Um, I don't know if I have it here or I might just have to, um, create, oh, maybe I do. Mm-hmm. Yep. There we go. I also demonstrated Copilot Studio, which some of you have used recently, some of you haven't, but it's all new. So Copilot has gotten smarter. It has a new level of IQ, both in the environment it's in, but also grounded in Foundry agents, in Microsoft Foundry, as well as Fabric agents. So there's so many cool things you can do now that you couldn't do even 6 months ago. And so if you tried Copilot at one point and then you're— and you were like, this is dumb that I have to tell it what's in my email when I'm in Outlook, you're like, why am I even doing this? It is way better now. So I encourage you, but it's not way better for everyone. They have tiers of support, tiers of licenses. So you do need to have, uh, the right license. And we are actually in the AI Leadership Institute, we are doing a Copilot Power User Masterclass tomorrow. So if you are in the AI Leadership Institute, I hope to see you there. If you're not, it's not too late to join. You're welcome to join us, but we will be doing a live masterclass, which we do every third Saturday of each month. And so last month we did a response, we did AI Leadership Essentials, which is basically like a broad brush on all the things AI. And now we're getting into very specific outcomes. So Copilot is tomorrow, and I can't wait. So one of the things I did want to share with you was, and I, I maybe y'all know this, but maybe you don't, um, is that there are so many pre-built agents and the list is growing. You can see here there are things that we used to build from scratch inside Copilot that are now commoditized, right? That are standardized as part of the development. Now some of them are available, some of them are in preview, but notice here the finance agent in Copilot. Allowing you to do financial reporting. Like, I used to have to do a lot of work variance analysis. I used to have to build this from scratch and then feed it in as an agent, a custom agent. These are now out of the box available. Some of the other ones, right? Like Citizen Services for all those state local governments. Like, this is not super useful by itself, but when you customize it with the, remember, business problem and the business requirements and the functional spec, right? When you train this model to specifically handle the challenges that your business is facing, these become huge accelerators. And as a matter of fact, that's what I call them. They're AI solution accelerators. If you have the right understanding of the problem, you can use this to accelerate a solution and start building momentum very, very quickly. Benefits. This is for HR. Uh, one of the first use cases that I worked on in most companies was like a benefits buddy. Because most people, uh, employees don't use the benefits that you're given. So a company pays for benefits and then you as the employee, okay, Noelle Russell Russell as the employee, don't use them all. Like I will not use all of my, you know, uh, employee benefits. I will not use all my healthcare benefits. I will, you know, cuz I'm working here, like I'm, I'm doing stuff. But if I had a gentle little nudging, not nagging service that was like, hey, you know, I saw you were on 8 projects this week and 62 meetings. I don't know if you know this, but you have free license to the Calm app. Maybe you should download it. Wouldn't that be cool? Just to give me contextual awareness of things I already have access to. So think about this. This is a baseline, you know, benefits agent that's gonna look at like your benefits package and, but it's the starting point. For accelerating an experience around building a better employee experience. I think it's pretty cool. So anyway, so I mentioned this as another opportunity. There's lots, of course, Google Vertex has its own agentic platform. Salesforce, oh my goodness, right? They have an agent force. It's an entire, the, the customization and personalization level of agent force is incredible, but they're pre-built. I always say like build on the shoulder of giants. Take these things. Marbleism was a quick and easy example, but like start with one of these and just start building. This is how I did what I did at Amazon is I just started building every single day. I built a new app every day, at least one a week, but I built over 100 in the first year, right? So that was at least what, 2 a week? But some weeks I'd built 6, some weeks I got stuck and had to stick with one for a long time, but I was constantly building. Not that anyone needed what I was building, though they did use them. Some of my skills are still very popular, but most of them, it was just sheer like experience building, getting receipts. I built a healthcare app, I built an, uh, entertainment app. I built like, I built all of these different experiences, but the most important thing I did is I built them based on seeing people asking for them. So I would monitor Reddit and be like, people would be like, oh, I wish Alexa did this. Bingo. I'm gonna go build it, right? Like there's so— people are, you all know this, people are very quick to complain about stuff, what isn't working, what's not solving their problem. But if you're like me, a solution architect, that's gold. I would love to, you know, vibe code or marbleism my way to solving a problem for someone where they are completely upset that they can't, you know, it's friction they can't get around. And I can unlock it in an hour for them. Like that I think is, is the dream. So being an AI solution architect is a big part of the ethos at AI Leadership Institute that when we see a problem, we actually look for problems so that we can start to think about how can AI help us solve them. So this is just another example, but I did want to show you my usage, uh, of marbleism. I don't know where it's gonna drop us, but you can see here this metric list is actually from February 1st to February 20th. And according to this, That's a rough estimate. It has, uh, saved me about 361 hours. The bulk of those hours were spent in building, helping me ideate on things that just take a long time. Content creation. I mean, some of you have seen this month, I am at, by the end of the month, I will have posted 1,000 posts across 7 platforms. Some of them are hard, meaning deep thought posts, some of them are quote cards and images. Either way, I was building kind of some momentum for my challenge. And so a lot of that is fa— you can see here, I actually have a standardized task for my agent to build me 5 new posts every day to give me an idea of things that I could bring to market. I ask it to find contrarian angles, which is like, imagine, you know, if you're in chiropracting, like it would be a, a, a post that says something like, If you're not in pain, right? Or, or if, if you think you're not in pain, that doesn't mean that you're not hurt, right? It's like very contrarian. Well, I, I think I'm fine. If you think you're fine, you might actually not be fine, right? How do you kind of shake someone's, uh, expectation or perception of something? And so I, in AI, oh my God, there's so many things that I could do. So this just helps prime the pump, get me thinking, and I posted, uh, exceptionally large increase in number of posts as a result of this. My blog post— and now, granted, I'm not a— uh, I had a format, so my campaign was to write one blog post, uh, a week for all of February. So it's week 3, and you— some of you have seen some of this content. Um, and then my legal assistant— so this is not— this is like a paralegal, not a lawyer, right? It's not going to give you legal advice. But I was able to take a legal document I had and review it. I I was able to find out if there were things I needed to update or change based on a new contract. I was able to ask questions about an NDA that was sent to me, a mutual NDA. And I was like, is this really mutual? Like, tell me what I need to know. Um, and, and again, you could do some of this in ChatGPT, but I liked it 'cause it's all in one place and they know about each other. I can ask my executive assistant to trigger a campaign with my social media and blogger. I can ask my sales associate to, identify a certain pattern of my audience that it's seeing. Oh, let me show you that. So this is my, my sales associate. These are the leads that it's collecting, and this is based on my ideal customer profile. But I wanna know, like, is there any, like, some of these New Zealand, Melbourne, right? These are different people, but they all have certain characteristics that I defined in common. Now I'm going to New Zealand. I'm going, some of you might know this, I always every year go to Australia. I'm going whether there's a South by Southwest or not. Um, but every year we go. And so now I will go in and I will have contacted them, and now I can reach back out to them when I'm about to head over to New Zealand, be like, hey, I know I contacted you before, just letting you know I'm gonna be in town. I've converted, uh, conversations at a much higher level when I just reach out to people once, and then when I'm actually going to be in town, be like, hey, I'm actually here. They're like, oh, Yes, I know you, let's meet. And it allows me to just have a much better human connection with people. So this doesn't do all of my connections. As a matter of fact, I'll show you the sequence, right? The first thing, well, this is very specific to this week, but I'm just like, hey, I've got a challenge you should join. We had 30% of the people that we, uh, contacted joined the challenge. Thank you, by the way. Glad you're here. Um, so thanks for being, uh, a part of the challenge. But this is a just a 3 email sequence to court them into hanging out with me and learning how I teach. If they don't come, great, that's it. I don't have to bother them anymore. But if they do come, then they get added to my, my, uh, CRM campaign right now. They're part of my customer base. But this is just an easy way to reach out. And most importantly, I did it all with my words. I didn't have to write any code. I didn't have to configure a CRM. I didn't have to build out an automation. I asked it to build me a flow. Granted, I'm a copywriter, so I gave it very distinct instructions on how to write the copy that's in those emails. You don't have to do that, but I did. Um, and so I'm very happy with the outcome as, as a result. So, so many cool things. Uh, this is what we did. So I went in and I literally built a whole new company in about 40 minutes. I had cus— like onboarded all the agents and they started working and I thought it was pretty cool. So I, the moral of this story is not necessarily It's not a product story, right? This is not like, go use Marblelism. This is more about how do you get muscle memory building with your words. And this is a wonderful— not only is it a nice environment to do it, it actually provides, if you're like me, immediate value. Like, thank you for helping me draft emails that now reduce the amount of work it takes for me to contextualize. I mean, it goes in and searches the entire thread. And contextualizes the draft to what I've already said to them. It's very cool. So I'm like, oh my gosh, this is so helpful. I have to change almost everyone, but not significantly. Like sometimes it will— I will tell you, my agent probably thinks this because I sound like someone who would do this, but it offers time on my calendar to everyone. And I'm like, thank goodness you don't send this, 'cause I would not be able to have that many meetings. But it like literally in everything, it's like, Why don't we just hop on a call? Which I like, cuz what that means is get in front of the human. It's very human focused, but at the same time I'm like, okay, delete. Uh, let me know if I can help you, you know, join my community. Uh, I do calls every Friday, right? Like, so, um, but I liked it. I like the idea that it defaults to humanity and, but again, it does that because I told it to. So my hope, uh, as I share this with you, share with you what we did in our challenge is that, and of course this is just one mechanism. We also had a bunch of people using like, um, Claude code, uh, like Claude now for example has, right? Opus is the default. Um, 4.6 has a million token context window. I don't know if you're excited about it, but I am definitely excited about it. So many times I'm trying to do something that I need massive amounts of context and I'm limited by that. I, I don't know if you've ever gotten that message, but I oftentimes will give it a bunch of context. It's like, you've provided too much data, the window is not big enough, and now I'm gonna get a truncated output from that model 'cause it doesn't know all the things. Or I have to take that information and break it up into PDFs and feed it into files and create a project. Like, it just adds friction. So a million is nice, but I also worry a tiny bit that when you get are given that much context, I, I, you just have to be really, really specific in what you want. So there is complexities that come with a large context window in getting a good answer. Um, so it's kind of like pay me, you know, pay me now or pay me later. Um, but I really like it. It allows me to do long-form content much more accurately. It allows me to feed in a lot of research that I couldn't really get in before. So, so I'm really excited. So we had a lot of people leveraging this, leveraging Claude code artifacts skills, was really, really cool. Couple of tools that I mentioned that I wanna make sure you all are familiar with because I use them every day. One of them is Hux, Hux AI. This is a tool based on the concept of NotebookLM and built by NotebookLM. Kind of, you can tell when you see it, it's got a podcasting approach to triaging your inbox. Um, your calendar and any topics that it detects as trends in your material, like in the stuff that's in your inbox and calendar. So it's been extremely valuable to me to help me do meeting prep, to help me set up for my day. Um, a lot of people ask me like, how do you have so much energy throughout the day? And I will say this is part of why I can handle my days with a lot of positivity because I know, I know all the things I'm gonna do. It's the first thing I learned. 6 o'clock, 5 o'clock in the morning, I'm listening to Hux. And it's like, it will tell me, you have— you are 4-way booked at 11. We should fix that. And some of you might be like, that gives me anxiety. But at least I know. In the past, I wouldn't even know that until 11, and I'd literally have to pick, who am I going to spend my time with, and what 3 people am I going to disappoint? Now I can handle that much better. And not just that, I get to learn about legislation. We had a huge— The EU AI Act had some really exciting stuff that happened over the weekend, but then also we had some big legislative stuff happen in the US today. And so I am not a newswatcher, judge me if you will. I don't watch the news, I don't watch TV, but I do like to know what's going on. And I used to rely on my friends being like, did you hear? And I would be like, nope, I'm glad I know now though. And Huxnow has closed that gap for me where I'm not relying on all of my friends to keep me up to date on the world. And the nice thing is I can tell it exactly how I want it to tell me that information. So I don't wanna know all the details, especially the ugly stuff and the dark stuff in the world. I wanna know what's happening, but I don't need to know the details. Um, and so I, it now does what I tell it. You can see here, it tells me how many emails I have, how many calendar meetings I have. It tells me the, I don't have it tell me the weather. I, uh, configured, I took that out, but you could if you wanted to, but it becomes your assistant personalized intelligence. If that's not what we're after, I don't know what we're doing right. Hyper-personalized intelligence experiences, whether that's for your business, whether that's for your employees, whether that's for your customers. So as a result of this, like I said, if you're listening to this and you want to like see these videos and the training that we did all week, you are more than welcome to join us. It is all recorded and we documented all the links to all the tools. It's a lot. I mean, people who were in the event, they loved it, but it's a— it was a lot of information. I also I leverage another fun tool. I don't know if I wanna do this, but I'm gonna put it in. But I use Read AI as one of my meeting summaries and I'm just transparent. Like, yeah, see, it's gonna show you. So it gives you this really cool number based on like your engagement and how good you are at keeping people engaged. And so I like Read AI 'cause it gives me feedback on how I showed up in meetings, in conversations, in calls. But fun fact, I talk fast. I don't know if you noticed. I apparently, uh, I talk fast. And so it would tell me, it would never give me credit because it would always ding me due to my rate of speed. And it would even say, you are amazing, everyone loved it, except you talk too fast. Um, and it wouldn't say you talk too fast, it would say your rate of speech is fast. And it gave me a number to indicate how fast words per second I say. Which is, I think, a bit impressive. Honestly, when I look at it, I'm like, wow, that's pretty good. Um, but that's me. That's who I am. I'm not alone. There are some really great inspirational leaders and mentors of mine that also talk fast. Thank goodness for AI, right? You can go in and slow me right down on the replay. Uh, but it does give me insight to how I'm coming across. Uh, lets me see sentiment analysis on the attendees. So it's, I liked it. So I used this technology to give me that feedback in a way that I never really thought of it before. Same thing with Teams. I don't know if you know, but Teams has a Speaker Coach in it. And in Teams, when you're talking, you can turn on Speaker Coach and it will tell you, I think that's another reason why I talk fast is that I don't have time for like ums and ahs, cuz I'm just like speeding like through the content. But, um, but it gives you a very like, a clarified view of how you're talking, how you're conveying the message, your story arc. Uh, so use that if that's interesting to you. But I am always looking for how I can symbiotically use AI in my day to improve my human, like the way I show up as a human. And so that's why I wanna encourage you that learning how to build an agent and marbleism, it will create a bit more muscle memory in how you articulate to an AI system how it can help you. But the only way for that to be really meaningful is if you actually have a metric to tell if it actually did help you. So establish a baseline, figure out what you're doing, and then use an AI system and see if it does better. There's lots of variables in this, right? If it does it better but hurts your customers, mm, not interested, right? If it does it faster but is not in your brand voice, not interested. So there's a lot of different variables that I'm always looking at, and that's one of the Things I gave to VIPs this week is my rubric, uh, for, um, measuring and asking. Maybe I'll pull it up. It's in Lovable. Um, so I built out, it's, it was a, a checklist at first. That's how I used to use it. And then, um, I ended up building out this, this checklist, but as a tool. So I'll just show it to you here. But it's like, you know, you're about to evaluate a tool. Like this is tool evaluation, model evaluation, and I built like a scorecard. Now I personally know like when it's in a certain way, like I can tell what number I'm striving for because I've put a certain value on ease of use over flexibility, over control, over cost transparency. So you could create your own metric. And I mean, just imagine you think about it, you write it down, and then you can build a tool that helps you do that level of metric, like analysis. I think we have the opportunity to add a lot more rigor to the way we do our work by adding AI systems that we design that are custom built for that. And that's what I'm hoping that this, this exercise did this week is how do we build an employee that actually helps us do what we're trying to do, not replace humans, but actually make sure the humans are doing the cool stuff, the fun stuff, the things they wanna do. And lots of organizations are reevaluating their HR processes to figure out, do I even know what my humans wanna do? And then am I building technology that amplifies that? Not a lot of people are doing it, but a lot of big companies with lots of employees, they're currently evaluating this. Um, at Amazon we had a practice, it was called like Skill AI or Skills AI, but it was something that as a, a AI, as a solution architect, we had to go in and say all the things we were good at. And it wasn't just like functionally, I'm good at AWS and I'm good at EC2 and I'm like, it was across the board. Are you a digital marketer? Do you play around with crypto in your spare time? Like it was asking all these things so that it could leverage an algorithm that when we got a client, let's say Unstoppable Domains, and they are like, we would like to talk to a solution architect, they could use that matrix to find the best human correlate to that client. And I feel like that, I mean, that was the olden days when it was just like a matrix we were managing. But like, imagine now, right, where we can index human capability and make sure that the right human is involved in the conversation and the solution and the development and the design at the right time. That, that's my, my dream, right, is that we all become AI leaders that know when and how to apply AI to amplify human activity. And it works. The 5% of companies that are doing well in this area, they've all realized that it's so important that human amplification is how you get, uh, a, I mean, even in decision, uh, even in like machine learning or, um, computer, like software engineering, we're still, we're not, um, immediately eliminating all humans from the software development process. We're just realizing Oh, we need humans to do certain things more than others. Now, software engineering is a whole nother thing, so I won't, I won't open that Pandora's box. But it's interesting to look at the future of work. What does it look like? It definitely looks like me trying to use AI to do my job better so I stay valuable either to my business or to my company, um, that I'm working for. So let's see. That's it. That's the recap. I hope you enjoyed it. If you wanna join us, you can. One thing I did offer, are, is that we, I do wanna mention before we jump into Q&A and break out for a few minutes, is that, I wonder if I have it here somewhere. I think I do. Lots of links here to, to, to do. I'll just pull it up. Is that we are back again. The AI Startup Studios is an accelerator that we run. It's called the AI Business Accelerator. I'll show you a little bit about what it looks like, but basically it's 6 weeks. And if you are a business owner or Let's say you are given a new opportunity to look at how to spend your time. You are in transition. You are thinking about, do I wanna become a consultant or do I wanna go get a job? Or you're in a job and you're like, wow, I'd really wanna know how to maybe build a side hustle that brings in $10K a month or whatever. In any case, there is an art and a science to building an organization that generates revenue. And so I, because of what we just did this week, I wanna show you how you can use AI to amplify your desire to build a company that brings in $85,000 a month is the goal, right? $1 million a year in annual revenue. And you'd be surprised, it's not as hard as it sounds, but it does require work. And I have noticed not everybody wants to work. So this is for workers. This is for people who wanna do the thing, who want to actually put their fingers on a keyboard, build something, serve, people, help people. Like, this is not for the one who wants to like coast on a couch. Not that that's bad. I sometimes do that on the third Sunday of the month, but normally I'm pretty active. I, I'm ambitious. I wanna build things. If I see a problem, I wanna solve it. That's what this is for, for people who see a problem and wanna build a company that solves it, but at the same time generate enough revenue that this becomes maybe the thing you do, or you can sell it. Exit strategy. So I'll just give you a sense of what it looks like. We do, of course, in-person learning. We are gonna do classes, but most importantly, every single week for 6 weeks, you get like step-by-step work you have to do. You're gonna build, uh, first thing you're gonna do is you're gonna identify what you want. You're gonna build a product, you're gonna build an offer, but then you're gonna build a lead magnet. You're gonna build a sales funnel. Some of you, that sounds foreign cuz you work for other people and that's great. Do not have to do this. This is not for everybody. But if you're gonna build a company, you are gonna have to understand what it takes to get an idea, get attention for that idea, turn that attention into a conversation, turn that conversation into a sale, turn that sale into investment of your revenue, and then start all over, over and over and over and over again with as much fun as possible. So that's why we did it. You can, uh, head over to the AI Startup Studios if you want to apply. It is application only. We can't— obviously this is a very, very hands-on program, so we don't— we have limited seats every year. Every year it sells out, but you're welcome to apply and see if it's something that we can work on together. And it's also a chance for me to learn a little bit more about you. It is 37 days away, not terrible. So that means it's at the end of March, my birthday month, and we will go through May. Mother's Day, uh, so that will give us a chance for you to really see what it takes to build. The goal, the promise, is that you'll have 10 paying customers, either your first 10 or your next 10, um, and you'll have a system that you can recycle to get 100, 1,000, or a million of the same kind. Um, so I think it'll be fun. So I share it with you so that you can see it. I'll throw the link into the chat, also throw it into, um, but it is pretty easy to find, AIStartupStudios.com. Uh, we've run this 4 times already. We do it once a year. This is your one chance if you want to learn from me how to grow a business to 7 figures. Here you have it. Um, it's usually fun because we become really good friends and we all build businesses together. And then you have to decide, do you hire people or do you not hire people? Um, but I'm, I'm here for it. I'm here for those conversations and, uh, really exciting, uh, opportunity for us to work together and and serve more people. That's my goal. So thanks for being here, uh, today. Now I would like for us to talk a little bit about this concept of AI employees, um, and find out what— how you all, like those of you who are in the, the Zoom and in the community right now, like this is a chance for you to meet people. Number one, the people in this community, there are very experienced consultants. There are also very experienced business owners. There are very successful people. There are also people on the earlier side of that journey. This is a chance to just meet people. And I find this world is small and people are, are where we are in our, you know, maturity. Uh, it goes fast. And so you never know, like next week someone, which just happened, someone could get a job at a company and you're like, oh my gosh, I always wanted to work for that company. Right? And now you have an in there. Um, I found out the other day that I've been friends with long time with someone I never knew where they worked. And then I found out where they worked and I'm like, I love that company. I would love to talk to the CEO. Can you make an introduction? Right? Like, so network is everything. Who we know and the human relationships we create is such a powerful part of our success. So that's one of the small things we try to do here in, uh, our Zoom calls are especially for the I Love AI community is give you a chance to talk to each other. I will encourage you when we break out for you to say who you are and what you do very quickly. We usually like to use a I help statement. Something like, I help your unique audience, right? Your ICP, your customer profile. I help this customer type of person do this thing with your solution, right? So what, who is it you serve? What problem do you solve and how do you solve it? The other side of that is you're like, I don't know. I don't know none of that. I'm here for, I need help. So the other option is that you actually can Literally say, I help statements. And you could be like, I need some help. I, but please let us know who you are and what you do. And then ask the question if you have, if you have some needs, there's lots of people here who can help answer. But I also encourage you, if you don't know the answer, it's okay not to know. That's one of my biggest, you know, success, uh, or, or biggest moments of education at Red Hat was how valuable they, they really, um, made or how much value they put on people who are brave enough to say they don't know something. Like, it's okay. Of course you don't know. Not everybody knows all the things. But if you do know, share it. If you don't know, don't make something up. That, that, that doesn't help anyone. Um, but have the courage to ask for help. And then if you know the information, feel free to reconnect with that, those people, either online, offline, in the community, outside of the community. But I think it'd be cool to talk about AI employees and what your, like, kind of what your initial feedback is. Like, Do you think, like, do you think we're moving in a direction where AI employees, meaning, uh, these resources in your business, like you could see on Marbleism, they named them. I wouldn't name mine. Like, I'd, it would be like EA social media cuz they're, it's not people. And so anthropomorphizing is maybe I just grew up on Alexa, so I'm ruined from that perspective, but I don't like adding human names to things that aren't human. So that would be my, That's my take, but I'd be curious to know how you all feel about this concept of like building these doer type applications and what concerns you have about it, or maybe what excites you about it, or both. So I think, because I, I'm just curious and would love for some of you who feel so moved when you come back to share some of what you've talked about in the room. But AI employees, it's an interesting one. I was surprised at how much resistance I got when I chose that as our challenge this year. I got way more excitement than resistance, but normally I don't get any resistance. Like, people are like, that's amazing, good job, I love it. And this is the first time people like DM'd me like, um, uh, I don't— do you know, like, employees are— we don't want AI employees, we want human employ— and, and I wasn't prepared for that level of kind of like resistance. So, so I'm, I'm obviously— I went through the whole week, it's totally fine. I feel like I defended my— the reason I think it's okay, but I think each one of us probably have a different perspective on that and come from different backgrounds and lived experience that might really create a more, um, a deeper understanding of why that's important and what to think about when we're positioning this either in our company or for our business, um, or as an employee. So, um, awesome. Thank you, Nitin. Um, Amazing. Uh, yeah, so just don't hit the button. Okay, so everyone else, let's do this. I'm gonna go break us out. Hopefully you will enjoy the conversation. It's a great time to meet people. I randomly make— these are random selected, so it's not like I'm putting you in rooms. I don't got time for that. Um, so I hope it's okay, uh, that just consider it the universe dropping you into a room with these people, uh, and, and take the opportunity to meet them. If you aren't gonna join, don't click the button, um, because You don't want to like go in and ghost people in the room, so just don't click the button if you're going to leave. Bye, so nice to see you. But if you're going to go into the breakout rooms, please, yes, don't join and then leave because then it like creates this weird box and I would be better. So let me go ahead and say, let's see. Oh good, and most of you are staying, so thank you for staying. All right, I'm going to go with— we have some bots in here, so I'm going to go with this for right now. I'm going to go with a bigger number. There we go. Oh no, smaller number. All right, here we go. So let me, it takes me about 2 seconds, maybe 15 seconds to, um, move the bots to the bot room. Uh, not that I have to, I don't think, I don't know how it works, but I'm just, this is Noelle Russell's like checklist of like, I don't want a bot and I don't even want to see it in there. I don't know if they can join rooms. Some of y'all might be able to inform me, but I find it's, it gives me a sense of comfort knowing that I, I went through this exercise. Oh my goodness. Wow. We got some great people in here. Cool. All right, cool. Thanks for joining me, Laverne. Nice to see you. Okay. Sorry. Getting distracted. Uh, all right everyone. Yes. Hi. Um, okay. See you later. I'm gonna click the button. I'm gonna give you about 10 minutes and then we'll close up for 5 minutes. And of course you have questions, we can stay for that as well. So go have fun. Talk about AI employees. Are you worried? Is it exciting? What do you think? All right. See you soon. And, uh, while they head off into breakout rooms, I'll just say to you all, thank you for listening today. If you want to come and meet people just like you who are learning AI, who are at different levels of their journey, who are doing different things with it, and just wanna be part of a community that truly believes in having the space to think and create and build and and do so without kind of the fear of, you know, it being okay to not know something, being okay to ask questions. Some of us don't have that psychological safety at work. Some of us don't have a community to ask questions of or learn or even share what we're learning. We don't have a place to do that. So I invite you to join the I Love AI community. We are pretty selective when you join. So please make sure if you join that you actually provide a profile, that you answer the questions, that you don't sound like a bot. If you sound like a bot, I'm gonna think you're a bot and I won't let you in. So, uh, if you sound like, you know, you're selling, uh, you know, stamps and you wanna get a bunch of people to buy your stamps, like, I probably won't let you in. So definitely if you are on the journey of learning artificial intelligence, I really, uh, have handcrafted this community to be a safe, interesting, fun, enthusiastic way to learn artificial intelligence. So I welcome you. Um, and I can't wait to see you. Don't forget noelle.ai, and hopefully we'll see you next time. Otherwise, I'll see you next Friday. Thanks for being here. Bye everybody. Hold on. See you later. I have to find the thing. Oh, there it is. Okay. Now for real, for real.