The 5 Must-Know AI Lessons for Marketers

I said, "Jeannie, would you please scare me half to death?" So why do I tell you that today? Well, it's not to scare you half to death. It's because I have 3 wishes for this presentation, and they are this: when you walk out of the room, one, you want to be an AI-first thinker. Number 2, you will know that your business will be disrupted, and not by one thing, but by many things that we'll talk about today. And number 3, stop resisting and start learning. These are the things we're going to go through. Now let me tell you a couple stories. We're going to imagine a future that is not too far away, so come with me please into this new world. So have you guys ever ordered a pizza, delivery pizza? I have. But imagine a future— and I've seen a startup that has this in alpha already— a humanoid robot coming to your house to deliver your pizza to you. You pay digitally with crypto, of course. And then what happens is not only do they deliver the pizza, but the humanoid robot comes into your house and serves you the pizza. He sets the table, he pours the drinks, and gets everything ready for you. But wait, you know, when the meal's over, there's something that happens that I really don't like, is you have to do the dishes. So my humanoid robot is going to select the dish soap and do the dishes for me. What a great idea, right? But now think about implications on your brand. If you're Coke or Pepsi and now the humanoid robot is pouring the drink, does that mean that you now have to market to the company that sells the robot? Does that mean you need to embed in a large language learning, learning model? What are the implications? If the robot is doing the dishes, does he or she pick that P&G product that's sitting there to do the dishes with? Just some things to start thinking about as an AI-first leader. But let's do another one, okay? In terms of disruption. Now this is quite funny. So the day before I came here, I went to have my teeth cleaned. And no joke, this is what they did. The first thing they did is the dentist said, "I have a new tool. It is tooth recognition." And as they scan my mouth, They were recognizing different teeth, and the AI toolset actually helped the dentist identify any issues in my teeth. But guess what? The dentist didn't execute on the, the device, the dental hygienist did. And so I was fascinated. I asked her, did you get trained on this? And she said, yes, I was so interested in it. I went and took two classes on AI because a dental hygienist's role is changing so much. Also think about the molds now. Now you don't have to go do molds. So if you're a mold company, now this is being replaced by artificial intelligence tools. And one more we'll do, and this is also a real startup doing this today. It's a company called 360 Fashion. These are mood jackets. Anybody ever have a mood ring? This is a mood jacket. What does it do? It senses your mood. They're being tested today in factories in Asia to determine, do you need a break every 15 minutes, or do you have a special— I mean, not a break every 15 minutes, a break every hour for 15 minutes. Or do you have a specialized way that you would need a break? It can sense when you're hot, when you're cold, when you're sad. But now think about this. If you're a manager now on the shop floor, do you now need to think about change management? How are you gonna execute this? How you do explain to factory workers that they're now wearing a jacket that's collecting very personal, personal information about you? All of these things are 5 years away, 10 years away from happening. And so everybody in the room, I hope one message that you take away from the session today is that you need to start resisting and start learning about this technology. Anybody know that Waymo is now here? They actually launched here at South by Southwest. I come from Phoenix and Waymo is already there. You can already take an Uber in a Waymo, which is an autonomous car. There's no one driving the car. It takes you from point A to point B by itself. This would be a great thing for you to try. And did you know that if you try it and you ride an autonomous vehicle 3 times, your consumer behavior actually changes? The studies show that after 3 rides in a Waymo, Uber, that you're more likely to buy an autonomous vehicle. Also, the price is cheaper too. Okay, so let's go, go to our 7 trends and I'll show you how these relate because you want to be an AI thinker. You know your business is going to be disrupted and you need to start stop resisting and start learning. And here's some of the things that you should learn. So I put them all on one slide real quick so you guys could take a picture. I will show it again at the very end as well, but I'm going to walk you through each of these 7 trends. There's so much to say and tell. I'll try to tell stories so that you really get it as well, so that you can come with me on this journey. So our very first trend Exponential, baby. Everything is growing at an exponential rate. Let me just flash through some trends for you that are not growing at a straight line, not growing linear, but are growing exponentially. So let's first take the, take the world of AI. 80% of enterprises say that in 2024 they will increase their use of GenAI. 80%, the highest of any technology ever adopted, adopted by enterprises. NVIDIA. You guys have heard about NVIDIA. They're now a trillion-dollar company. They do the hardware that's behind all the AI. You can see here that the shipments are going up and look who they're going up to. Microsoft and Meta, Oracle, Amazon. Into your house. So we can see that delivers your pizza. Takes crypto for a payment, actually doesn't leave after it delivers it, but it comes into your home. It sets your table, pours your drinks, and then when you're done— my favorite part of all— it cleans up and does the dishes too. And only when it's done with all of that does it exit your home. Now, of course, this is in probably pre-alpha stages, but I got to see it, and I started thinking about the vision if this comes to pass. First of all, for example, since the robot is pouring your drink, who does Pepsi or Coke now sell to? Do they sell to you? Or do they sell to the robot company who's making the choice of what drink they pour? If it's doing the dishes, who does P&G now market to? Does it market its soap detergent to you or to the robot, the humanoid that's doing that for you? And even the architecture. The company was telling me that some of the homes, to get a humanoid into the home would have to be redesigned. Do the standards for doorways need to be enlarged so the humanoid could get through? So many interesting things here. But let's think about not just in the future, but what's happening today and what really matters. I told you about those 1,500 companies that I surveyed. Out of the 1,500, 20% of them were doing what they called successful AI. Now, you probably read the MIT study that said 95% aren't. So let's say between 5 and 20% are today being successful. If my daughters brought home that score, I would still say that's an F, right? So what I wanted to do is look at those people who were doing it right. What did that look like? Well, there were 3 things that they had in common. One, they first look at the data. Data's not sexy, but you need good data because good data coming in gives good results coming out. The second thing all of these successful companies were doing is they were focused on business outcomes, not technology for tech's sake, but business outcomes. And then third and final seems so simple, but yet was so impactful. They were doing change management, making sure that the humans in the organization were ready for whatever tech change was coming. So let me give you a couple of examples. NTT DATA, one of my very favorite folks that I work with, they're out of Japan. One of the things that they're doing is they're training everybody on AI, not just a certain set of departments, but everybody. In fact, what they do is they have this level system. You get a white belt or a yellow belt. If you get the yellow belt, then there are certain tools that are opened up for you. If you get the black belt, then you have access to all of their tools. Making sure that they're training the people, the humans, not just the machines, right? Like my title to my book, AI First, Human Always. And then this example is one of my favorites. One of my friends invented this mood vest. Do you guys remember the mood rings? You guys have the mood rings? She had mood vest. And so using IoT sensors, it senses your mood. It senses how you feel, your temperature. Super cool. She sold it to this manufacturing company in Malaysia, and they're doing all their checks. It took 9 months to get the tech ready. To take anonymous data in from the readings of the Mood Jacket into an overall AI dashboard so they could adjust the temperature, they could play music, they could get breaks. It was all gonna be good. But they only took 2 weeks to do the human training. Now, many of you are probably already thinking, "Okay, I know what's gonna happen here." And it did. Only a couple of managers decided to train their people and explain what the jacket was. They told them the data would be anonymous, it was used for their good, for breaks and for not tracking productivity. But not everybody did that. So the groups who didn't have that training actually rebelled when the jacket was rolled out. They did not want their data tracked. They didn't know the data was anonymous. So the company had to actually retract the whole project for a while, while they trained the entire factory. Now, this is really important, I think, because, you know, if you think about the humans involved, there needs to be humans in the loop as you go forward. And the big trend today that was just published by McKinsey said that we're spending 10 times the amount of time training LLMs as we are in training our people. So we need to make sure that we're doing both. So at the end of this section 1, I believe that success in AI will depend on those who are focused on business outcome. Think, you know, best customer support, not best customer support AI agent. So they have to be focused on that business outcome and they're experimenting, but then they're scaling. All the articles today are about scale, experiment and scale. There's companies that have hundreds of experiments, but they're not scaling yet. And my advice is you need to grow your people because people are the ones who grow your business as well. So let's go on to point number 2. So we've got number 1, we're gonna make sure we're leading with vision, vision that's focused on business outcome. On number 2, I wanna talk about trust that goes into that core system. The myth is AI can't be trusted. But in reality, AI can be trusted if— and it's a big if— if it's being used responsibly. Used responsibly. And all of us play a role in that. So one of the first places I went to look about trust was the Edelman Trust Barometer. Do you guys know that? I've used it in almost every technology trend that's come— cloud computing, mobile computing. And this is the latest report they just released on trust of AI around the world. Unfortunately, they didn't give me Australia, otherwise they would have put Australia up there too. But you can see at the very top, India and China, top of the board for trusting AI. Who's at the bottom? My country, United States, and Germany. And then I just got back from Brazil, who's right there in the middle. At that kind of breaking point, and it's Brazil that has that kind of critical moment right now around trust that's coming about. Now trust is so important. How do you keep trust and how do you break trust? Well, let me show you how you might break trust. And this was actually part of Edelman's study, showed that this was a big trust buster worldwide, but in particular in the United States. So on the, on the right-hand side, you see the Pope, who's unfortunately no longer with us, but he was wearing a Balmain jacket. Balmain is a very fancy brand in Paris. In fact, that jacket would be about $10,000 US. This picture went viral. 10 million people reposted it, commented on it, etc., around the world, and then the Pope felt like he must come out and say something because he wanted people to know that he would never spend $10,000 on a jacket because he was all about helping the poor. So people's trust were broken. Many, many, many people had experienced this and shared it. They were part of the sharing of something that wasn't true. So how do you know that was true or not true when it was posted? And then on the other side, you can see me at a a red carpet event in Los Angeles. It was cold, not supposed to be cold in Los Angeles, so the only jacket I brought on my trip was a puffer. I don't know if you guys pack a puffer because it packs so small. So I wore my puffer over my fancy dress. A reporter saw it, took a picture of me, and decided to go outside its bounds and start calling me on the phone like Open Claw had done before to many of its owners saying, "Wake up, it's time to get going. It's kind of start your day. Because of the security issues, we're now seeing lots of claws come to be. So Ironclaw was just released by Near, and Near rewrote the open source code in Rust to correct a lot of the security issues. You can now order a pizza safely and trust the agent with your credit card with Ironclaw. We also had Jensen that just released Nemo Claw on Monday at the NVIDIA GTC conference. I was trying to play around with it a little bit. That was just Monday. Today is Wednesday. There's Pico Claw and there's Zero Claw. So make sure you check these out. Don't leave South By without a to-do on your list to go and check out these tools and these ways to predict agents. Here's what I predict. In 18 months, your LinkedIn will no longer have skills on it. What is it gonna list? It's gonna list your agents. That's gonna be the way that you showcase your capability. So now, to talk about the power of agents, I wanna tell you another story. This is actually a true story. At Stanford, a professor was teaching a class there, and he said, "I'm gonna give every team $5. You're gonna have a week, and in a week, I want you to come back, and I want you to tell me how much money you made off of your $5. Everybody understand the assignment? You guys all with me? Yeah? Okay, that's the assignment. So a set of teams decided, oh, we're gonna go buy water and candy and food and we'll sell it on site, and they had a return, a small return. Another set of teams, what did they do with the money? They invested in advertising and they mowed lawns and cleaned houses and even did website development. Development. They had a better return on their money. But let's listen to the last team. They realized that the most valuable asset they had was that 15 minutes their professor had give them to talk in front of the whole class of Stanford graduates. And what startup wouldn't want that ability? And so the second group sold their time. Instead of they present, they sold— it's pretty smart, right? Some of them sold the whole 15 minutes, some of them sold groups of 5 minutes that each of them got to be up there, and they actually were the winners of the contest that the professor started. And why is that? Because they understood that the value was not the $5. That really wasn't the asset. The asset was time in front of the Stanford students, and that is where companies are winning today. It's not companies who are spending more money on artificial intelligence. It's companies who are rethinking what business they're in and then writing agents to help them in the new business capability. Does that make sense? I think it's a really important point. Okay, so that's number 2, agents. We talked first about people. We're going to sandwich kind of people in here. Talked about agents. Now let's talk about teammates. So now you've got these agents and you could treat them as a tool or you could treat them as a teammate with you. They're autonomous, right? They're agents. They're autonomous, so they can make decisions on their own. So let me show you some of the teammates I have. This is one of my favorites. I posted about this on LinkedIn. I developed a shadow board. This is my advisory board, and I put Reid Hoffman in there. I had one company that I founded called Silicon Blitz. Reid does blitzscaling, so I got to meet him and talk to him about that. Jeff Bezos. I used to work at Amazon. I presented to Jeff, so I know a lot about him. Warren Buffett. How can you have a board without Warren Buffett on it? Of course, John W. Thompson, if you don't know, he's the former chairman of the board of Microsoft. He is my mentor. And then I have a skeptical VC and then a domain expert. What's a domain expert? Well, remember I told you I was working on my customer service agent? It might be someone who's experienced in customer service. So, you know, when you're at a startup, it's not like working for Amazon where I had thousands of people I could ask. You're kind of lonely. And so now I have an agent, my teammate, my advisee, my shadow board that I can ask questions of, and it does really well. Now, if you remember last year, if you were at my presentation, you guys asked me, how do you keep up? How do you keep up? And I was so proud to show you this board. These are all my lists and all my social media. And I said, I've got it in one place. And many of you said, Can I have that? That is so cool. Well, today I want to let you know that we actually have something even better. I worked with Robert Scoble, and what we have now is we have a summary of all of that. I have a teammate who provides me a briefing every morning and presents this report to me, analyzes over 50,000 news articles that come in overnight. So it gives me the top 5 pieces of news, the investor news. Um, I can choose a category and say, oh, I really wanna focus today on NVIDIA 'cause they just had their big conference. This is a teammate that advises me as well. Okay, I'm gonna do just a couple more. I love this one. I did this with a Notebook LLM. So I write for Forbes and I want to be able to get the pros and the cons. But again, it's just me, right? I also, for this presentation, I wanted someone to debate me 'cause I felt so strong about humans, not technology. So I have two agents. You're gonna hear the creepy agent voice, you know what I mean? So just bear with me as you hear the creepy agent, but here's what they can do. You guys can do this too.

Welcome to the debate. Everyone saw that, you know, from pilot to payoff presentation at SXSW aimed at executive leadership. Yes.

Sparked a pretty massive conversation.

Exactly. A massive controversy. Is enterprise AI success stalling because leaders basically refuse to restructure their organizations, or are we just waiting for the underlying models to get powerful enough to force their hand? I fall squarely on the side of leadership. Human alignment and, well, org structure dictate whether these initiatives survive.

Right. And I have to push back hard there. It's the sheer power and paradigm-shifting capabilities of the new underlying models that actually force those organizational changes in the first place.

Look at where enterp—

So these two agents will go back and debate back and forth. And if you guys want, just email me, I'll show you how to do it. It's super easy. You can create your own debating agents. So you can say, I'm gonna play me and I'm gonna have my boss play the other, so I'll know all the scenarios that'll happen. That's kind of cool, isn't it? What'd you guys think? Yeah, it's kind of interesting. Okay, so now I'll just do one more for you. So this is, so I use Claude. Claude doesn't pay me to say this, but that's my favorite AI tool that I use. I know I see some Google people in the room going, hmm, I gotta talk to you afterwards. But I do love Claude, but I was working on skills. So, you know, you train skills for your agent and I wanted to build better skills. And so I asked them and they sent me this 33-page guide of how I could build skills. I don't know about you, I don't wanna read 33 pages. Anybody wanna read 33 pages? No. So what I did is I fed that into Claude Cowork and I built an agent to walk me through how to build skills. So this agent now, instead of me having to figure it out and go through 33 pages, walks me through it. I was gonna share that with you guys today. The URL is at the bottom. If you guys wanna create Claude skills now, you don't have to read the 33 pages. You can just use my teammate here and get those skills. Now, last year I presented 100 tools that I love. I'm going to present today. This is my 25. You guys liked this last time, so I wanted to provide this to you again. I know there's lots that I left off, but I just wanted to do my top 25 so you could see it. So now you saw a lot of my favorite agents that are out there. So how do I use those in my daily business? So let me show you my new org chart. By the way, this was also, uh,