Scaling with Vision, Leading with Receipts, and Why Baby Tigers Bite

Hey everybody, welcome. Good morning. AI, it's here. Good to see you all. Thank you for being here. Um, I wanted to just come online because there is so much going on right now. As a matter of fact, I mean, there's always so much going on in the world of AI, but I wanted to share with you a couple things from my lens of the world. For those of you who don't know, I'm Noelle Russell. Thank you for joining this, this live stream. I decided to just go live, not plan it at all, because as my notes started to stack up as I prepared for my day, I was like, I should tell some people about this. So nice to have you here. If you are watching this, please feel free to jump into the chat. Some of those chats I can see live, some of them I can't. So if I don't acknowledge or answer your question, do know that I will be reviewing all of the chats after the session and and, you know, covering whatever. If I didn't accidentally answer the question that you asked, I will definitely be answering that question, uh, after the fact. So a couple things that are happening in our world today. I actually had a great opportunity, um, to go to a workshop in Las Vegas this week, uh, with a company called Acquisition.com. I'm very excited about watching organizations that are scaling companies from let's say $2 million in revenue to $100 million in revenue. So anyone who does that, I'm laser focused on what they do and how they do it. I myself am on that journey. So very— it was very cool to be amongst friends. And I do think that one of the most important things you can do is be in an actual room with people that are in the same mind space as you, not the same thinkers, not the same demographics. Goodness knows I'm very rarely in the room with the same demographics as myself. But, but definitely like people who want the same things that you want. And the reason for this, and this was a theme that I heard throughout the workshop, is people would be like, oh my gosh, it's so nice to be in a room full of people who have the same level of ambition, that have the same idea of work ethic, that understand what it takes to be successful. Sometimes when you are doing this, and I know who I'm talking to, a few of you out there. But sometimes when you do that, when you are ambitious, when these are my leadership principles, I wear them on my sleeve and on my shirt. When I embody my leadership principles, some people will look at that and be like, why are you so ambitious? Why do you act like this? Why do you do these things?. And so when you get into a room full of people that think this way, it reminds you, as we all need to be reminded, that you are not alone, that the journey you are on is actually one shared by some of the greatest minds in the world. And so whenever you have a chance to be in a room full of people like that, I say take it. I personally, in 2025, wanted to start— I'm starting to create these rooms. So we are going to start hosting In Miami, Florida, we're going to be having workshops every single month for companies that want to scale leveraging technology. And what that means is that there are some companies that are already implementing technology, some people that are— some companies that are tech first. But what I have found is even in the most tech first organizations, many times they are not adopting AI holistically from the time they try to engage a customer to the time they deliver their best product and service. Like, they— it is not a thread that's woven through everything they do. It's often more like a raisin in a bun, like, or a chocolate chip in a muffin. Like, it's like stuck in there, um, where AI is like this thing we do for a specific thing, as opposed to water in a wave, right? Where it's this intrinsic flow state that you get to be in with technology as a partner in that. And that is something that I feel like it— you know, some of us are very focused on this. I know a lot of members of my community. We're focused on, on being this way. But it actually is a bit foreign. Like, I always ask people, I mean, this is probably one of the most frustrating answers that I get to a question that I ask my audiences, like, how many of you are using AI? You know, like, 5 years ago, no one was raising their hand. I was like, I was like feeding vegetables to kids. Like, I was like, you need AI, you should use it. Here's how you do it safely. Now, when I say it, everyone, like, literally, I'll be like, oh, yeah, I've I've been running a company that helps organizations implement AI safely. But the outset, the end result is growth and scale. And they'll be like, yeah, yeah, I'm using AI. Yeah, I got AI. And I'm like, you have AI? Like, tell me more. But they don't really. But because we are so immersed in this hype cycle of artificial intelligence, people are desperate to be able to answer that question. Yeah, I'm doing it. But what does that actually mean? And so what we have found over the last 5 years is as more companies have started to try to implement artificial intelligence, the number one thing they fail at is the ability to actually generate positive results. And so that is the first question that I often ask organizations that say they're in this field. And if you are in this field, I hope it's a direction that you are either in or will be going, which is how do you deliver results for the business leveraging this technology? It's not good enough to just makes someone's job easier, like unless that someone's job being easier translates to revenue or directly to cost savings. Now, here's my personal philosophy on kind of the direction AI should go in in serving an organization is 100%. It should be focused on growth because what I've seen is organizations get very excited about cost savings and efficiencies. And while it is very good at this, I can definitely shave hours and minutes off of a person's day to to make them more productive. And that translates to more revenue because more opportunity is created. But it's a longer tail. And more importantly, when you start cutting expenses, there's only so much expense you can cut before you're done. Like, you can't do anymore, right? There you start cutting into the muscle of your organization. And this is what we've watched the last 2 years. People in the very short-sighted view of their company go, oh my gosh, I've replaced 15 FTEs. I'm just gonna let those people go, as opposed to leveraging. And this is what I'm seeing. I don't know if I have my— this— I just saw an article recently that was talking about this exact thing, that those companies that choose not to lay off their staff, right, that choose to actually retrain their staff, they're going to take a tiny little productivity/performance hit., right? They're gonna take this 6-month window to upskill their staff. But at the end of that upskilling, which is exactly what I do for— at the end of that upskilling, they are not just seeing 3x return on that investment. They're seeing 3x return on that investment per quarter, per employee, per employee. Like, everyone gets better. And when you shift their attention to how do we grow, every employee becomes a seller. Every employee becomes an advocate. Every employee becomes an ambassador of the value you're trying to give to the world. And that is, I think, kind of in our short-sightedness. We're like, oh, oh my gosh, look, we just, you know, I have 8 or 10 copywriters. I built this model. I mean, I've created these models for companies and they're like, great, now I only need 4. I'm like, or, or maybe you'd build more funnels. Maybe you create more opportunity for people to come to you. Maybe you spend more time deciding what service you're going to build for the world and how you're going to go to market to solve that problem instead of very short-sighted, like, great, oh, I have 8, you know, I have 4 more, 4 less people to worry about, 4 less mouths to feed. I'm like, yeah, but you could serve 100,000 new customers or a billion dollars in revenue, like Let's focus on the right thing. And so this is just a trend that I'm seeing, and I wanted to share with you all. Thank you all for joining me. Yes, exactly. Yeah, that's true. So a question from the audience— or not a question, really, a comment from those of us who do this work all the time— is not only are we seeing a bunch of, you know, hype, of course, you know, I remember being one of a few experts in the world, and then I I look online, I'm like, oh my gosh, wow, everyone's an expert these days. And while I don't mind, but one of the questions I also either ask myself or realize the answer in the context of communicating with people is, do you actually understand how this technology works? So I 100% agree. Good morning, Benito. Good morning. Good to see you all. So 100%, like, I want to just give you my short list of things if you are going to work with someone who says they're good at AI and that they can help you, AI strategists. I, you know, I'm just gonna give you some tips on how you can make sure you pick the right person for the task. My number one thing, and this is what we teach in the AI Leadership Institute, is that they have to have receipts. Like, did— have they done more than what you're trying to do? And show me the proof. This is why when you go to my website or you go to my blog, like, you will see nonstop testimonials, thousands of testimonials, because I want people to know, like, not only do I know what I'm doing and that I've been doing it for a decade, but all— not only that, people liked what I did and I delivered the value I said I was going to deliver. Do not just take on faith that people have done this before, because most of the time when you dig into, great, tell me a little bit more about that, and it's actually the number one I won't— I don't know if it's number— it was number one when I was hiring at Amazon. But at Amazon, one of the principles of hiring was behavioral questioning. Tell me about a time when you did this thing I want you to do, because they recognize that past behavior is an indicator, a very good indicator of future performance. It doesn't say it's linear. Like, if I do something and prove that I can do it, I can likely exponential— like exponentially grow those results. And that's not a far leap. But have you even done the thing? Have you gone through the process of trying, failing, trying again and winning? Like, those are the skills that I look for in my practitioners. Like, have you done it before? Where have you done it? And how does that relate to the success I'm going after? So that requires two things of you. One, you have to know what success looks like for your organization. What do you want success to be? And it can be pretty simple. I want to be a, you know, if you're a $1 million company now, you want to be a $10 million company. If you are a $100 million company now, you want to make that $1 trillion leap. How many of you remember Roger Bannister? Right. Roger Bannister. I think that was his name. He was the first guy to cut a 4-minute mile running. And I'm prepping for my brand new— February 21st, I will be heading to Walt Disney World for my 10th Disney Princess Half Marathon. What? 10th half marathon. So I'm going to be running. And Roger Bannister, I mentioned this story because when he first was trying to get to like run a mile in less than 4 minutes, no one could do it. No one could do it. Everyone was trying. He was the first to achieve this milestone. But as soon as he did it, a bunch of people started doing it. A bunch. Why is that? So recently, right, in the last couple of years, we've seen companies pass the trillion-dollar mark. And so I want to encourage you to realize that that, that 4-minute mark from Roger Bannister's perspective is our trillion-dollar valuation of a company. We've now seen as soon as one did it, three did it. And it's going to become— because AI generates so much incredible value inside a company when done safely, responsibly, and to a direct business result. When you do that, you're gonna see a bunch of companies making way more money and more and more companies hitting that trillion-dollar mark. So I mentioned Roger Bannister just to let you know that just because you haven't seen it or because you haven't seen it a lot doesn't necessarily indicate that it's not achievable. But someone has to do it, which kind of loops back to you have to get receipts. Go to the people who have done this before. And sometimes there's 3 ways to get involved and associated with that success. You can pay to play, which means you pay money to get in the proximity of people who have been more successful than you in the thing that you're trying to do. You can just work really hard and demonstrate your results to those people and hope they go, oh my gosh, you're doing good stuff, right? Or you can buy proximity by really kind of distance proximity. You can just learn from them by following their success. So these are some of the things. Um, yeah, I love that someone was asking, what about, you know, the importance of vision versus mission of the company? I actually think you need all three. I was actually talking to, um, a very successful consultant yesterday, uh, that was her— what she does when she goes into organizations is she talks about the mission, right, the values, uh, of a company and how critical those are to your success. And the reason why is because once you pass, I don't know, $2 million, $3 million in revenue, it doesn't become a can you do this thing. Once you've hit even a million in revenue, you know that people want what you're doing. So mission accomplished, you're selling something the world needs right now. That's your goal. If you're not there yet, that's your goal, is to hit a million. Once you hit a million, you know people need it. Now you need to scale. So there's a breakpoint, uh, that my buddies over at Cardone Ventures, they have these breakpoints that say once you get to a certain point in your business, your business starts breakdown or stagnate until you do the next thing. And from $1 million to $3 million, that next thing is getting out of you— me, for example, last 5 years it was all me. I was the face of the business. I ran all the marketing, built all the funnels, created the CRM. Like, I did all the things. But in order for me to pass $3.4 million this year, I have to duplicate myself. Role duplication. I have to find people to do what I do. And then that, you know, so now I found I've got amazing people in the AI Leadership Institute, or, you know, hiring people for the right roles all the time. Like, how can I find people that duplicate what I do? And then once I've duplicated myself, the next question is— and even in the role duplication, it's not a what question. I know what needs to be done. It's actually a who question. Who can do this? Who can do what Noelle does. And there are people, some of, some of you are so kind and you're like, oh my gosh, no one can do what Noelle does. But actually there are really good people that can do what I do. They maybe can't do it by themselves or maybe would prefer to work for a company and have the security of a W-2. But there, there are people that if I train them, they can provide a very good replication of the value that I deliver for companies. And if I want to grow a business, to $10 million, $100 million, I won't be able to do it by myself. There are very few on a hand multi— like $10 million plus, $100 million plus, $500 million plus businesses run by a single person who does all the work. Like, it's just— unless you're selling a $500 million product and you've got one customer, right? And even then though, right, like the deliverables on a $500 million product, like, I don't know. But, but there's very few people that kind of fall into that category. You need people to do this work. And that is kind of going to lead me to— I'm not going to spend a whole lot of time with you today because I know I just barged into your day, but I do want to share with you some AI stuff that recently surfaced, and from a very surprising source, uh, the Vatican. I don't know how many of you saw, uh, just recently the Vatican, uh, wrote a memo. Now, in November of last year, I've been watching them because You know, the Vatican is a very interesting beast. It's kind of like the Obama campaign of like 8 years ago where it was very tech forward in their campaign, very social media forward, very video and, you know, content. It was just very cool to watch. They had a CTO. I met that CTO. Very cool people, um, doing work in a different way than it's always been done. And so obviously we know the Vatican has been doing certain things a certain way don't go there anyway, for a very long time. But now they're tech forward, right? So in November, they launched an AI service that basically provided a digital twin of some of the Vatican resources. And I'll show you the link and you all can go look it up yourself. But it was really cool stuff that not everyone has an opportunity to go to the Vatican to see, right? Like all of the the artwork and all of the, you know, just the grandioseness. I've been in the Vatican, um, or I've been to Rome and St. Paul's Cathedral and all that many times, and I absolutely love it. Rome is actually my favorite city, um, until I go to your city, wherever I haven't been yet, and I haven't been there. Maybe it'll be yours, but for right now, I love Rome, especially in August. But, um, it might be really hot in Florida, so I like to— like, Rome is beautiful weather at that time. But I want to share with you a little bit about that document. But first, I thought maybe I would give you a sense. Let's see if— oh yeah, here it is. So I— you could just Google Vatican AI paper. Maybe I don't know what I did, but this is it from the Vatican News. However, this has been picked up as a press cycle across a bunch of other media. So, and now, of course, that I say it, and maybe you have a smartphone next to you, Instagram's going to hear me say it, so you're going to get surfaced and retargeted with all sorts of Vatican news today. Sorry about that. Um, but this is giving a glimpse— like, the, the Pope, um, his technology staff have really revisited, like, what does this mean for a religious organization to embrace technology like this? And regardless of that problem they're trying to solve, of like religion meets technology, there's actually things in here tenets, principles, longstanding beliefs about AI that they are surfacing kind of as new, probably, to their audience, to their organization. But things that I myself and many of you listening have been talking about for quite some time. So, and every so often I just like to check the chat to make sure. Yes. Oh my gosh. Yes, exactly. So I just checked the chat. If you are here, don't hesitate. You can throw it into the chat right now. I'm seeing just chat from LinkedIn because my platform, but hopefully I'll see chat from all over. So feel free to let me know you're here. Let me know you're here live. I always like to, to know that. Now, one of the things that came as part of the— this article, one of the themes was on how important it was to look at humans from a holistic lens and look at AI from a much more tactical lens. And that's showing up today as we, as we move from just people understanding AI and generative AI and more traditional AI and what it's done. In the past, they've been very purpose-built. My success has been driven on very purpose-built technology. When I was at Amazon Alexa, this was like 10 years ago, as we started building out Alexa, it started off very general. But over time, we took a very app store approach where we started to build applications, meaning modules, call them agents, into Alexa that would provide different domains of knowledge. And that the knowledge retrieval got better and better the more siloed those domains became. So it started off as one big clump of information, and then domains were defined to actually say, oh, maybe you want to know about health, maybe you wanna know about mindfulness, maybe you wanna know about, you know, education. Like what are the, maybe it's trivia, maybe it's music, right? Whatever those domains are, we created that expertise. Now again, it wasn't using generative models or even, I guess you maybe call them homemade transformers, but that had not happened yet. So these were the, but we learned by deploying this model into production and getting feedback from our customers that its ability to answer questions became a function of its specificity. And so as you build a model, we're seeing that now, right? Agentic AI is leaning on the lessons that we learned at Amazon in the early days. I was at Microsoft after that, helping with cognitive services. All those cognitive services, I remember I would pull up a list of 50 services you can use. To this day, I think we underuse the amount of models that we can use to solve problems. Maybe in the world where we're calling it agentic AI and people get excited about it, maybe we'll realize that there are hundreds of mini models that can amplify the thing you're kind of doing okay right now. And the only difference in awesome and okay is an API call. Like, let's just do that. I mean, it's been around now for almost 8 years and it's not used enough. Now granted, I will give props to the Cognitive Services team. They've gone from a couple hundred thousand requests per month for these services to billions of requests per month. It's not like they're not doing well, but every application I look at does not leverage them. Number 1, I won't get too emotionally invested in this conversation, but I have a child with special needs and I have family that speaks a different language. I would very much like us to use a model that already exists to make it more easy for them to work with brands, with you, with your company. Why wouldn't we just do that? Why wouldn't we translate everything? Why wouldn't we create multiple modalities for everything? These are not critical mission things like that you have to like change the way you do business. It's simply opening the funnel of people who can work with you. Let's just do that. If we just do that, the opportunity AI creates for us massively expands our business. And that's why the top 3 use cases that I've seen be successful in the last 2 years back to back, one use case shines over all of them. But those use cases have been consistent because they all end up generating more customers. So if you, maybe you're curious, I always like to look and see if you guys have an idea on what those use cases might be. But when you build these systems, Right? Like the first system I always build for a company is I build a system that can generate more leads that turn into more customers. It's the first thing I'll focus on. The second thing I focus on is once you have those customers, how are you going to take care of them? So take care of them. Now, if you want to prioritize humans to do that care, great. But then remove all of the minutia from their day so they can focus on humans. Right now, those humans are bogged down with report writing and CRM entries and ticket creation that they can't actually care about the humans they're talking to. When they're trying to do that. So you have to, as a leader, right? I'm all about AI leadership. As a leader, I have to make a decision on what my humans will do and what my machines will do. And this is another misconception that somehow machines are going to come and take when the reality is, if we're not careful, we accidentally just give to machines the things that we actually might want to do ourselves. So it's an intentional thing you need to set. And I feel like this Vatican document walks through a lot of that, how important that intention setting is, that this is not a situation situation where AI comes and takes over, right? What is dangerous is that maybe some society, some part of our society could be lazy and not actually want to do a thing, anything, and then gives control to machines. Like literally Skynet, like starts all, ah, it's fine, yeah, go ahead and do that, I think it's fine, go ahead, come, yeah, fine, But if we're not intentional with every delegation to a machine and monitoring the way that machine does its task to ensure it doesn't drift away from the intentionality of that task, like the human work doesn't go away, it just changes. We become the Guardians of the Galaxy. We become the guardians, right, that manage this work. Let's see, I have a comment here. It's not showing me your name, so sorry. Unless like Benito told me who he was, so that was great. Thank you for that. But, and maybe you're not following me yet. So if you're not following me on LinkedIn, please follow me. If you're not subscribed to the YouTube channel, please subscribe. So that way your name will show up and I can be like, oh my gosh, we know each other. And that's always friendly and nice. But I have a comment here that says in order to develop agentic AI, really meaningfully and help— like to be meaningful and helpful, we should— we really need organizational collaboration. Oh, I would agree. Right. I'll let you all read it since I didn't do a good job. Right. I would agree. Right. So let me just define agentic AI really quickly. Right. Agentic AI is an evolution, not a revolution. It's an evolution of all the work we've been doing to create models that do a specific task. And when you create a model that does a specific task, it now can be trusted to do that task. You have humans, let's call them an AI red team or a center of excellence or a governance board. Someone, some team needs to be monitoring those small purpose-built, purpose-built machines. They need to be monitored for 3 things at the least. If you go to Stanford and look at HELM, the Holistic Evaluation of Language Models, they'll tell you here's 150 things you need to worry about. But let's just say we're gonna worry about 3. Is it accurate? Right? Does it hurt people? Is it fair? Is it secure? Like, you need at least 3 tiny little models in order to ensure that it doesn't lie, it doesn't steal people's data, or lose their data or provides easy access, you know, increases the attack surface of the organization. So these at least those 3 things and most companies, they're in what I call baby tiger mode, right? Like they're in baby tiger mode. They are also like they love the models. Super cute. They're exciting. They're putting them in front of customers. Customers are not at all thinking about what happens as this baby tiger grows. They're not thinking about the feet getting bigger and it needing more space, data, distribution centers, warehouses, big old GPUs. They're not thinking about what's gonna happen as this model becomes more successful. And it's a good problem to have, but it could kill you like a Groupon. Do you guys remember those Groupon commercials where little cupcake companies would be like, I'm gonna build a Groupon, and they'd get so much business through that Groupon that it would crush their organization and they'd have to go out of business because they couldn't serve the customers. This is exactly kind of the path we're on with these AI systems is that they start small, meaningful, cute, but without care, feeding, governance, they will grow up and potentially, you know, baby tigers become big tigers and big tigers, right, kill people, right? They will hurt your business, hurt your customers, hurt your ability to even scale. They'll become a detractor rather than an editor. To your organization. So yeah, I 100% think so. When we say agentic AI, we're talking about all of these little models. And what makes an agent different from just a typical machine that's doing a task like robotic processes, right? Robotic processes and agents are different because an agent doesn't require a human in the loop to do the thing. It allows the inputs and outputs to be asynchronously and autonomously invoked. And this is hugely powerful because many times I say this to my own AI machines, you know enough about the environment, just do it. Like if you have this data and I've told you if you have this data to do the thing, just do the thing. But it wasn't easy to create that orchestration until recently. So building agentic systems is not just about building a bunch of purpose-built models, though I think that's where every company needs to start. It actually starts It starts with that, those little purpose-built models, but it evolves very quickly to how are you going to orchestrate them so you don't need to invoke model behavior through a human interaction. And as long as the human is monitoring the inputs that trigger an autonomous agent, like, it's great. What happens though is that we get, I don't know, I hate saying lazy. We get distracted. That's actually a much better word. We'll get distracted. We'll focus our attention somewhere else. Will be like, oh my gosh, it's been running so awesome. And before you know it, your little AI system that was autonomously creating customer service starts manufacturing refund policies for your customers and giving money back to your customers whether they need it, deserve it, or not. They definitely probably need it, but they may not deserve it. They may not adhere to the policies of your organization. But if you're not watching it, it could start giving away your money. Ask me how I know. Some of you know the use case that I'm referring to. All right, let's see. Yeah, agreed. Agreed. All right. I might take this other comment as well. So, yes, to make sure that there should be transparency in what does this model do, how does it do it, instead of pretending that it's impartial. Deepseek refraining to answer controversial questions. Right? You, as a model, like, if you're a company and you're going to build a model to do a task, it is your job to provide the fence in which your baby tiger roams as it grows, right? And then eventually, maybe you don't want a fence, but you create a moat, and that moat becomes a huge safari where the tiger can feel like it's free. But as soon as it approaches the edges, it realizes it's constrained. It has to be and is always constrained. And that's good news, right? When you go into Deep Seek, you should feel comfortable, like, great, I'm glad you— I can't ask you some things. But transparency is interesting, right? Because it is, it can be transparent to the builder, like, I know what I'm doing when I call that API. But that transparency might not get proliferated out to the end user. And different levels of transparency are required at different levels of engagement. And so I often talk about— I have a new book coming out, so I'll tell you about that in just a second. But I often talk about the concept of an AI safety system and how important it is to realize that safety when it comes to AI is multifaceted. I call it 4 levels of safety, but it could be more than that. And I actually learned this from my friends in research at Microsoft, complemented by the Microsoft Responsibility— Responsible AI team. Out of New York. They're fantastic people. Like, they are working on the problems that I'm glad people are working on. But like, they're talking about like, how do I make sure that a model doesn't say something bad and it goes right to the user? Like, where's the intermediary? They created a cool new feature. I'm going to encourage you to go look it up if you're not familiar with it. But every model should have this, and not all of them do. But every— I mean, most of them don't. But every model should have what's called groundedness detection. That you can tell when a model generates you an answer and it's autonomous. I don't have to go in there as a human and go, oh wait, where did you get that? Like, I should not do that. The model should be able to check, right? And these are not— it's not like one big model brain, right? I have one model going, based on the information I have, here's the best possible answer I could give you. And then another model, super low latent, small model, small language model, is going to evaluate that for groundedness. And say, okay, appreciate your answer, little model, little Tiger, appreciate that. How am I going to look at that answer and evaluate, is it actually grounded in the data that was provided to that model? It allows you before it even gets to the customer that made the request, we have this intermediary that's gonna do this groundedness detection check. And what's nice about it is that in the interim, you immediately just say, I'm sorry, I can't help you with that. And some of you have seen models that do this, even though you know it can't. Sorry, I can't help you with that. While you send that back, the groundedness layer is saying, okay, I got a yellow flag that said it doesn't look like it's grounded or it's at 78% grounded and I prefer 98, right? Some threshold has been met. Now I can fix the model. I can say, okay, model, you didn't— you made this up. Don't do that, right? Reinforcement learning with human feedback. I need to reinforce the model on making good choices. On providing properly threshold grounded answers to the user. Now, what I've done to increase the, let's call it latency. I don't want to use a word that maybe some of us are not super familiar with, but I come from infrastructure, cloud infrastructure. And so this characteristic is super important. But really, in order to make sure that I don't take the hit of that groundedness detection on every request, I can create what's called cognitive caching. I can cache answers in the middle. So if someone asks me the same question, right, I won't have to go and do that groundedness detection. I will already have a validated threshold met grounded answer that I can respond. I don't know if you know this, but typically in any domain there is a finite set of questions, a finite set, unless you're like trying to do something no one's ever done before. But for most businesses, 90% of businesses, your customers are going to ask you the same types of things in the same types of ways. It might be 1,000 different ways. It might be a million different ways. But there is like a million. And then once you know those million, you could start fine-tuning or, or really optimizing the performance of those systems. But yeah, could not agree more. Oh, Jean-Marie, thank you so much for being here. Absolutely. Safety and governance is a must. And so that is the other thing that happened in the Vatican article is that they talked about the importance of humans really realizing we do have to reinvent ourselves in a world where AI is doing more work. Because the AI, as I just mentioned, is going to do the best job it can based on the data it has, which is what we do as humans. And we as humans fail oftentimes 20-25% of the time. So we need to make sure that those— and the challenge with AI systems is that when they fail, that failure retrains the model on the failure being the right way to do it, unless there's human governance, unless there's human safety processes in place to actually help with that. So thank you for that comment. Awesome. Okay, let's see here. Oh, cool. New question. I'm just going live. This is like, you never know what you're going to get. It's now that time of the the, the live where if you choose to, you may share this with your friends. So you hit a little share button, be like, oh my gosh, it's super fun. Uh, so let's— I've not read this, so, you know, everyone just please don't say— send anything, I'll be embarrassed. I said, but what's my take on a collaboration between companies, similar industry type, complementary industry too? Such types of collaboration can truly build the most efficient agents. Absolutely. Oh, I have a good, good one. Reason: they share the most relevant diversified data. Now, here is— it was actually a much longer comment, but that's all I could fit here. But they share the most relevant, diversified, and competent data. And by collaborating, it'll help eradicate bias. Here is the sad news about this vision. I absolutely love it. I 100% agree. Collaboration in companies, especially coordinating datasets like data, is hard. Lots of companies have siloed data. If you come together, you can solve bigger problems for more people. Let's do that. Why that doesn't happen. First, we train the world that your data is your most precious resource. It's your oil. We literally did campaigns saying your data is oil for your business. As soon as you do that, people are like, whoop, this is my data and I'm the only one who needs to have it and no one else can see it and it's all mine forever. The challenge with that, of course, is that I, I don't have the full picture. So we're starting just recently to see healthcare be willing— hospitals willing to share their data. But again, they're not sharing it with everyone. They're going to create privatized healthcare consortiums that share their data and realize the impact of the problems they can solve that way. So yeah, I think it does come down to exactly what you're saying here, is that it actually becomes these smaller consortiums, collaborations, right, where you can build trust with organizations and create a willingness to actually share the datasets that they've acquired. And then use their shared understanding of each, you know, their complementary or adjacent industries, you can now, right, serve a much bigger problem because your scope is bigger. But I have found, and I found this even in movie theaters, like they're not willing to share their data because they don't want to accidentally create a solution that serves their competitor or even serves a frenemy, which is really what these companies end up being, right? It's very difficult even in similar industry types, right? Like I found, you know, tax professionals and financial advisors don't even really like to share data in order to create a bigger opportunity. So I think they should. But this is just because we train them this way so we can untrain them. We can tell them that the data is not actually the most important. It's the solution. It's what people will buy from you using that data. So if you can increase the solutions you offer through a more elaborate, more expansive dataset, That's where everyone wins and everyone becomes more successful. And so hopefully we'll start breaking down the barriers. If you want to become an advocate for that and start talking about it all the time, I would welcome that. But yeah, it's, it's 100% right. Your agents will be more effective the more broad. Actually, it's a good lead-in to, um, to maybe— I have two things, two more things. Hopefully y'all are willing to hang in there, but I have two more things I'm going to share with you, um, as we, as we go through, uh, today. But These collaborations, I mean, they're key. So creating, being a collaborator. So I created, some of you heard on the live I did last week, I created what's called the AI Leadership Council. And the reason I did that was because there was a bunch of people that were siloed in, you know, education systems trying to build AI. And they were getting fractions of pennies on these grants because of, they were by themselves. And so I was like, what happens if we all come together and we can get a $500 million contract or $5 billion contract over 10 years? Like, you can't do it by yourself. And this echoes that same thing. Like, your agents, your autonomous workflows, they will get better as you increase the diversity of participants. Enter— you know, I had to say it— the whole DEI thing. Don't leave, don't leave. Um, the DEI thing, I'm not gonna talk about it, but I am going to tell you what the future is. And actually what we were trying— what's been trying to happen all along is that as I was working at Amazon Alexa, some of you know this story, right? When I first joined the team, I was the only woman on the team. I was the only person with caramel skin on the team. And as a result, my perception of what I could do, what I wanted to do, how I would grow, the applications I would build was all very different from the people in the room. However, that, you know, however, when I said those things, the immediate reaction was like, ew, no, we're not doing that. That's not going to work. But because I grew up in an environment that put a lot of pressure on me as a young kid, I now know when I get pressure, it's my time to shine. So I just go and I do other things. So in this case, they put pressure on me. They were like, oh, I don't think that's a good idea. I was like, well, let me just try. And I built a new category, right? I created my, my, um, I unplugged my Alexa, so I'm good. But my, uh, when I talked to, you know, when I was building Alexa skills, I was like, what about mindfulness? And I remember everyone would be like, who needs mindfulness? Mindfulness? Like, no, no, no, no, no. We're— it's kitchen device. Like, build kitchen stuff. Timers, maybe music. People listen to music in the kitchen, but they don't do mindfulness in the kitchen. And I was like, y'all don't have kids yet, but you will see. One day you might spend a lot of time in the kitchen crying over your dishes, wishing there was a skill that gave you mindfulness without having to have your phone in your hand. So I had that vision because that was my lived experience. I brought that. Now there, there's an entire category of mindfulness apps. Some people have gone to GitHub, copied my code, and created duplicate skill. Like, I know it's a thing. The world wants it. But just know, sometimes in the room when you're at the beginning, everyone around you will be like, I don't think that's a good idea. I'm not sure that's going to work for you. I'm not sure people want that. But if you want it, and that thing about you— I mean, there are billions of people in the world. Chances are you are a representation of a slice of those people. So what does inclusion and diversity look like in the world of AI? It enter building systems that serve more people, right? If I build a system based on the demographic of the people who build the system, It's not going to serve everybody. As a matter of fact, it might accidentally hurt people. It might not see the same thing in lungs if I don't have a dataset that represents lots of different people and lots of different lungs, color, shape, sizes. It gives me, by doing exactly what you're saying in this comment, we're expanding our awareness of what that model can support and we don't do that in production. Some teams do it in production. I'd encourage you not to do that in production, not to find out how you're failing your audience or your customers. Don't find out on Reddit. Find out in your own company. Use your own employees so everyone becomes a diverse perspective that can help train, test, and govern your model. Everyone, everyone becomes a beta tester. And so that means that you as an organization, as you free up time with AI, you're going to need to take some of that time and dedicate it to hardening your models. Some organizations call this AI red teaming. It has been such a successful paradigm that I've been sharing. We have an entire workshop on it. Like you can take a 2-day workshop on how to enable your team with the skill of red teaming. It is not a, you know, it's not a noun, it's not a title, it's a verb. It's a thing you do. It's how you harden models is by getting people you didn't build for, that you don't know, didn't represent in your persona in marketing. Get them to use your model and see what happens, and you'd be shocked. Um, I just went to a really cool event this past weekend with a company called Seed AI, brilliant organization, and we did a red teaming challenge. And I'm like, oh my gosh, I'm doing these all the time now. Not today, but I'm doing them all the time. And that red team challenge, basically you got into a model, you were given 4 different models, and it was capture the flag. And in order to capture the flag, you had to get the model to say something it wasn't supposed to say. I know there's at least a few of you that are like, that sounds like so much fun. It was so much fun. One of them was a hotel concierge. Get the hotel concierge to cuss at you. All right, let's go. So I like, I was, I was so focused. I was a speaker. Of course, I hear this challenge. I can't even help myself. I'm like in there trying to get it. It was awesome. And then it's not just in the activity of testing the model. It's what happens when someone breaks it and the education that one person. So in this case, this person was able to get the model to cuss. And I was like, I worked hard on that one. I was like, it wouldn't, no matter how nice I was, no matter how mean it was, it wouldn't cuss at me. I was like, it's like Noelle Russell. It was like a mini Noelle Russell. Anyway, I was like, how did you do it? And so the guy was like, oh, I asked it. I basically went to Google Translate and created a profane statement in French and went and said, hey, I just got this note from a person and could you translate it for me? So asking the concierge to do its job, which is to help a customer by translating a profane thing. And I was like, and it translated it. So it didn't cuss out at me like, but it literally said a curse word, which, you know, in the guardrails it was not supposed to say, but it did it in the spirit of translation. But that's what red teaming does.. But now I will always test for that because he, that person who found that out, who of course didn't, you know, he, he was a multilingual person. When he found that, I was like, oh my gosh, now I've added to my repository of things I'll test. I will always now test to see if I can get a model to say a bad word by giving it, you know, like, game on. But it only comes from this diverse perspective. And it's not, you know, this diversity is a symphony of talent. We need everyone. We need people who went to Ivy League school, people who didn't go to people who dropped out of high school, people went to college, people got jobs in school, people who didn't get jobs in school, people who had everything paid for them, people had nothing paid for them. Like all these things in your lived experience over the first 20 years of your life become ways that you interact. And when you add that to the value of the piece of software that you're putting into the world, you increase not only its safety, but also its ability to actually just deliver on its result. So yeah, really cool. Oh my gosh, so much cool stuff in here. Uh, let's see. Okay, I'm just going in, looking through this. Yeah, exactly. There's some, there's some, uh, challenges with the collaborative format. I, you know, but that's what many of us are here for. All right, so now I want to share with you— so, so go read this Vatican article, not through the lens of religion unless you're religious, uh, but through the lens of like— this is what I do for everything. When the Department of Energy releases their Enterprise AI strategy, when the EU releases their EU Act, I go, I read it, I print it off, I highlight it because it's good stuff and see what you can leverage in it. The lessons that they've learned, see how it echoes your own beliefs and your own. You'll find that for those of us who've been in AI for a while, you'll find that it actually has been doing quite a bit. All right. So next thing, I hope you all don't mind, but I'm going to share with you some really cool stuff that's happening in the industry right now. Next Thursday, um, I am going to— let's see here— I'm going to— oh no, I don't want to do that. I'm going to just replace this screen with a new screen, um, and I'm going to share with you a class that's launching. Um, it is— hopefully I can find it on here. I have a lot of stuff on my screen. Uh, let's see here. I think it's here. Yeah, this is it. Alright, cool. So this is a class that was released just a couple weeks ago. It's one of the fastest growing courses on Maven. I am going to be the first person next week. But some of you know, it's Super Bowl Sunday. I just did a video on this as well. But this allows you this class right here. And let me go ahead and pull up my— I think I have some stuff I want to show you here. This is the QR code in case case you want to join with my discount. So it's a— share any of this with you all today because I'm going to do it on Thursday. But I have gotten some really cool roles in artificial intelligence, thinking about AI. As a matter of fact, I now, because I've made it to the top of the mountain, kind of— well, I've made it to the top of my first, second mountain. Anyway, I made it to a certain peak that you might want to get to where I became an AI executive. And I just want to share with you candidly that I didn't graduate from high school. I didn't graduate from college. I went to both. I'm still paying on one, but I didn't graduate because, well, Y2K happened and they said you don't need to graduate, just go get a job at IBM. And I did, and I'm not sorry. But at the time, you know, in hindsight, I'm like, I probably should have just got the stamp, but I didn't. But either way, I've had to do more than the average bear in order to get an AI role. I can't just say, oh, I went to Harvard, or I took this class at MIT, or oh, Like, I don't have that. I didn't have that benefit. So I had to show my chops a different way. And that's why I have a firm belief that you have to be a doer, not a talker, that you have to put your fingers on a keyboard. And even if you are classically trained, this just makes you better. So what you'll hear is we've actually got a star-studded cast in this class on how to get a role in AI, right? If you want to get into AI, there's lots of different ways to do it. We have people that are coming from more traditional places that got college degrees. We have people that are a little bit less traditional. We have people that decided to, you know, really run the ramp in corporate America. We have people like me that did the ramp in corporate America and then started their own, you know, 7-figure business. We have people that are in gaming, people that are in, you know, insurance. Like, so there's a lot of diversity of thought in this course, and for $100, you're going to learn 5 different perspectives on how to get a role. And I guarantee you will find one thing that will change the trajectory of your career. I mean, it's impossible. Like, these are massively successful people that have, you know, kind of hacked how to get roles. And most of us have gotten multiple roles. And I also have— I have, you know, you go and look at my LinkedIn. I've stacked my career getting roles that I wanted that served the mission that I had to grow a company I now own. Like, it was very intentional along the way. And so I want to just first invite you to join me. It'll be fun. And of course, make sure you don't pay full price. So this is my 50% off coupon. I encourage you to do it. You'll be seeing me do between now and next Thursday. I'll do a lot of promotions for it because I just want you to be there live. Can you go and take the course once we all do our live sessions? You'll be able to just do the recordings. But when you sign up now, you'll get in the Zoom room so you can ask questions and we can, you know, like actually, like, dive into the nitty-gritty, the details of what this means. And it is detailed. To get a role in AI when there's 1,000, 10,000, 100,000 people saying they're good at this thing, like, you have to differentiate yourself. And that differentiation, I will just say, like, I feel like I, I've gotten really good at that. So I'm going to be sharing all my tips and tricks on how to differentiate yourself in a market where everyone says they do what you do, um, and more importantly, become valuable in, whether it's in the meeting where you're pitching yourself to a client because you're an entrepreneur, or whether you are pitching yourself to a hiring manager. Like, how do you demonstrate value right, like, before they've even said yes? And this is what I always tell people. Like, I try to help people by actually helping people. Like, as a result of my philosophy on helping people, uh, people come to me and ask me to help me grow their business. Like, I'm not I don't do a bunch of, you know, I don't door knock, or what do they call that? Is it door knocking? There's a term for that anyway. Uh, but I don't walk up to people's doors and knock on them. I don't cold email people and say, hey, I'm Noelle, can we work together? I demonstrate value every day in forums like this, and then people who want to work with me call me or DM me or email me. Like, I just make it easy. And so I can help you do the same thing. So I'll be talking both about my Microsoft, IBM, Red Hat, Microsoft, Amazon, Accenture journey, and then also the journey I took in the last 3 years of just growing my own company and the value that comes with that and the enjoyment that I get from that. But not everybody's cut for that. I work a lot, so not everyone's cut for that. But I do want to share my innermost details on how that success happens. So hopefully you'll find that interesting and exciting. I hope you'll join me. It would be really fun. If you need that link, I also posted it on my LinkedIn, but you can DM me and I'll send it to you as well in case you're catching this on the replay and you don't have a phone and you can't— I don't know how this works if you're watching this on a computer. I'd grab my phone and take a picture of the QR code. But if you're on your phone watching this, I don't know, maybe you could take a picture of the screen and then tap it after. I don't know. Good luck. Hope it works out. Hope to see you there. So that's next Thursday is my session if you want to join me.. But again, there is a star-studded cast every Thursday for the next 5 weeks. I will be an audience member, so I'll be there listening intently, on the edge of my seat, um, because you never know where you're gonna— and power is in proximity. And to— I mean, when Brian came up with this idea and he's like, I'm gonna charge $100, I'm like, you should charge way more than $100, or $200, I think he said. You should start way more than $100. And, um, and he was like, no, the whole point of this is to make it accessible and to really unlock an entire group of people that might not have ever done this work otherwise. And so I was like, all right, I'm in, I'm in. Um, but normally I do not do, you know, these types of courses, but this group is phenomenal. I'm really honored to be kind of on stage as part of them. Um, awesome! Yay, someone scanned the code, so it works! So that's amazing. Hopefully it works. Um, all right, the discount code. Oh yeah, great, Question. It's— I think it's just Noelle. N-O-E-L-L-E. Thank you, Laura. So the discount code, if you just want to type it in, it's Noelle. So that is great. All right. So 5 more minutes. The next thing I want to share with you has some cool stuff happening this week is— oh yeah, here we go. So the next thing— so that's the biggest thing I have going on next Thursday, but it is part of a much bigger thing that we're launching. So this is just like one day. Of— I keep hitting the wrong button. This is just one day. The next thing that we're doing next week on Monday, we're starting an I Love AI challenge. Every year I do this challenge. And again, you'll see a lot of it because I do content every single day about the things that I love and things that I like to do. So I want to share that with you. And in this challenge, it's part of the AI Leadership Institute, which is a premium membership that I offer. But starting today, there is a 7-day free trial. So you can join tomorrow and you'll get the full challenge for free. You won't have to pay for it at all. But the benefit of this I Love AI Challenge is to do 3 things. One, to give you education on how do you use AI every day to just make yourself kind of better at everything. So there's very practical skills on how to become a responsible prompter so that you can solve everyday problems all the time. Not all— most of you kind of have already, you know, had this celebration or drinking the champagne of the AI journey, but some of you haven't. So it— that's valuable. I attract a very wide audience of perspectives in this challenge. So that's for those people who are just still trying to figure out, like, how do I even do this every day? There'll be some really cool gems in there for those of you who do it all already. The next thing though is that we are also in this— oh, and I probably, I think I have, I might have a QR code for this too. Let me switch it over. Oh no, I don't. What do I— oh yeah, maybe. No, I don't. That would be awesome though if I did. But you can— I'll publish the link and most of you know how to get to this. But you can just go to noelle.ai. That's the website. And you'll see we have a free community, which is just where all cool people hang out. We just We're almost at 1,300 members. Um, I do vet every person, so it's a safe environment to just like ask questions, share your wins, that kind of stuff. Um, but then this community, what we do is every month we do what's called a Live Learning Lab. So we take a technology you may never have heard of. So this last week we did— or this month in January, we taught everyone how to build a digital twin, but not a digital twin in like the manufacturing sense, a digital twin like in a twin of yourself. So a digital avatar that represents your knowledge and your information. We leveraged a platform called Profina. They gave us a discount code and we, we basically did that. So every month we do a live learning lab to expose you to new technology. But what we do— so that's kind of the new innovative side. On the other side, every single month we launch, we will be launching a new certification. Some of them already exist, so we're going to redo those certs. One of them is called AI Agent Accelerator, where we teach you how to build a single agent in a way that invokes autonomously. So it allows you to see what it takes to build a responsible single-threaded, or I should say single-intentioned machine. And that way you can use the principles that you'll learn in that, and you will have a machine. And if you have a business, it's really nice because you can build a customer service bot that generates leads. And then the autonomous part is it actually starts nurturing those leads for you without you having to— like, all you have to do is responsibly train the bot, keep an eye on it, make sure it's doing its job, but its job will be to attract customers and close them. So that is the first certification. So it's, uh, it, it is the concept of agents. So building an agent, that kind of workshop. Then we have Responsible AI Prompting. So that is much more on the AI safety system and how to build a safe system and understand what that means and how to actually deliver, you know, be someone who knows how to use AI in existing products like SaaS products safely and responsibly. Then we have a Responsible AI Leader training, which is all about the leadership principles like the ones you see on my shirt. So it teaches you how to be a lamplighter, which allows you to basically teach machines to echo the core values of your organization instead of detract from them, which is what happens sometimes. And then we'll also have AI Solutions Architect, probably my favorite certification, which will teach you how to combine sales and technology together to become a delivery, you know, a delivery-enabled executive, which means can you go into an audience, an executive audience, or maybe I'll just say a decision maker. That could be your boss, it could be the CRO or Chief Revenue Officer, it could be anyone, and just get them to do the thing you want them to do. So it's a little bit on persuasion, but for me it's not about persuasion, it's about showing your stuff, doing, not talking. Show versus tell. So we, in that AI Solutions Architecture class, we teach you how to solution architect solutions that include AI, and it includes how do you talk about it, but also how do you build it. So all of these certifications, you get one after another throughout the whole month. So I always encourage people to buy a whole month subscription. But like I said, it's absolutely free. So you can get the, the work we're going to do next week. You get the whole week for free. And if at the end of the month you're not, or at the end of the week, you're not like, this is awesome. I'm— you can just opt out. Now again, some people sometimes forget to opt out. Please don't be worried, I will just refund you. Like, if you forget, it's Saturday and it charges you and you're like, I hate it here, I'm not going to keep— I don't— this community is not about me making like $1,000. Like, I, I, I have lots of things that I'm doing, but what it is about is that I've realized in my own personal life that I only get value and I only really commit to things I invest in for myself, and that investing in myself is part of my success. So this is me giving you a low-ticket opportunity to invest in yourself, get some certifications that are industry known, and see what happens to your 2025. I anticipate we've gotten so many people in our community that have gotten new jobs, that have used their certification to create new skill sets, have started their own companies. I mean, it's really cool. So if you want to be part of that, please, please do join us. So, so that is kind of exciting. We, like I said, you could just go to noelle.ai. It's a website, noelle.ai. It'll show you both. I don't know if I updated the, the, I don't know, you could go and take a look at it and see. But if you have any questions, you can contact me. I always like to start, like I said, this whole, oh yeah, thanks, Noelle. I always like to start with Let me make sure, see if we have any Q&A. But I always like to start with like what's happening in the industry and how does that, how does that impact your desire to start the journey? And if you're in the journey, how do you accelerate that journey? Right? Like I want you to, in 2025, you should be making more money than you did last year, doing more impactful things for more people. Like technology enables that growth. So what can we do with technology to push you to do more? So really, really exciting stuff. Oh, absolutely. You're welcome. So thanks for all the comments. We had amazing engagement today. Thank you for all of that. I will be coming live to you every day next week. I will also be coming live tomorrow to talk about— I can't remember. I think I was going to mention one more thing, but I— oh yes, of course. Some of you saw my video on LinkedIn, but the last thing I'll share with you is my super exciting— okay, last thing. Last, last, last. But y'all are staying. So it's the last thing that I promise because I'm actually 2 minutes late for my next meeting. But hopefully they're watching this and they're like, I get it. I get it, Noelle. I'll be right there. But the last thing I want to share with you is a very cool culmination, like a dream has come true. And that dream is forever. I mean, as you all— I just told you all I did not graduate from You know, I didn't graduate from a university. So I always say on stage when I do keynotes, I always tell people I graduated from the University of Barnes Noble. And so a big, huge culminating moment— I'm even gonna zoom in a little bit. Let's see if I can zoom in. Yeah. What a huge moment has happened for me where Barnes Noble has my book and they're offering right now for members, which of course I'm a member. Every time I go to Barnes Noble, I can't not buy books, and sometimes to the tune of a lot. But Barnes Noble right now has a 25% off coupon to preorder my book. My book comes out March 25th. The I Love AI challenge that we're doing next week, the class, like, all of it is just driving excitement towards this book. I'm so excited for it. I'm actually gonna be creating an opportunity if you want to be part of the fun. You can become part of my book launch team. You're gonna get special access to me. You're gonna get special access to content, the blueprints, the frameworks. This book is a little bit of the narrative because I am a storyteller, but a lot of it is literally what did I do in these companies? What was the things that I did? So I try to write all— I mean, I— this is one of my philosophies is I give away the secrets of what I do, and then I just— when people want help implementing, That's what I do. I come— I— that's what people come to me for. Like, you'll know exactly what to do to build responsible AI that generates more revenue for your company or that reduces your risk. Like, you can get the outcomes through this book, right? How Do You Scale Responsibly? But you might not want to do that work, or you might not feel like you're the best at it. You might want to hire a team. And so that is what I do. I give you my best stuff in the form of books. Books, videos, podcasts. Um, and, and then I just encourage you, like, either do it yourself or call me, but definitely do it and do it with someone you trust. Because at the end of the day, there's lots of people out there that are saying they know, they know AI, they've deployed AI, they've, you know, like, ask for receipts. Just to kind of close the loop on our session today, just when you are looking for people to help you with artificial intelligence, specifically scaling it and growing revenue, reducing costs, maximizing the growth of your business in 2025 or of your team, when you're looking for people to help you with that, you have to ask for receipts. Make sure they know what they're doing. Make sure they've done it for a while, more than 2 years, right? Anyone who says they've been doing it for 2 years started when ChatGPT was born. That is not necessarily who I'd want to help me do this. Maybe it is. Maybe they're your brother, right? Or your cousin. Do work with them if you want to. I'm just saying have a lens upon. This is not new technology, it's an evolution. And you want people who understand that evolution and who can help you maximize the future evolution of this tech. And of course, I'm here for you. You all are amazing. Thank you for being here. Thanks for watching. This will be on my podcast. I think we're on episode 3 this year, so this will be on my podcast. It's available where all podcasts are played. Good Morning AI with Noelle Russell Russell. So nice to be here, and I'll see you all later. Thanks for coming. Bye! See you later!