Adopting AI for Business: Where to Begin

Happy 2026. Hey, how's this for a New Year's resolution? How about you Master AI Marketing? The AI Business Society gives you live training with experts who are getting real results. No hype, no fluff, just proven workflows that work. And you've got access to an incredible library of training from many of the incredible guests that you've heard on this show. So secure your AI transformation by visiting socialmediaexaminer.com AI welcome to the AI Explored.

Podcast, helping you put AI to work. And now, here's your host, Michael Stelzner.

Hello, hello, hello. Thank you so much for joining me for the AI Explored podcast brought to you by Social Media Examiner. I'm your host, Michael Stelzner, and this is the podcast for marketers, creators and business owners who want to know how to use AI. Today I'm going to be joined by Paul Raitzer, and we're going to talk about how to adopt AI for business. If you are just getting started, this is the episode for you. We're going to get into all sorts of fascinating things you need to think about as you begin adopting AI. If you're brand new to this show, be sure to follow us on your favorite podcast app by clicking that follow button so you do not miss any of our future content because we've got some rock stars coming up you're absolutely not going to want to miss. Let's transition over to this week's interview with Paul Raitzer, helping you simplify your AI journey. Here is this week's expert guide. Today, I'm very excited to be joined by Paul Racer. If you don't know who Paul is, you got to know Paul. He is a marketing AI expert. He's the the CEO of the Marketing AI Institute and founder of the Marketing AI Conference, also known as Macon. He's the author of the book Marketing Artificial Intelligence and co host of the Artificial Intelligence Show. And his courses include Piloting AI and also his newest course, Scaling AI. Paul, welcome to the show. How you doing today?

It's great to be with you. Looking forward to the conversation.

Well, for those that don't know, Paul's been on my other shows. He spoke at my conference. He closed our keynote at Social Media Marketing World, and Paul, I think, is the guy. Now, before we get into what we're going to talk about today, which is really how to begin adopting AI in your business, I would love to hear a little bit of your story, Paul, because not everybody knows how you got into AI. So start wherever you want to start.

So some People may know me from my past life of running an agency. So I started and ran a marketing agency called PR2020. We were a HubSpot's first partner back in 2007 and I really grew my agency in those early years on, you know, the change that was happening in marketing, social, certainly mobile, was obviously a big one. And then just the whole idea of content marketing and inbound marketing, like creating these lead magnets that could help grow businesses. And that's what we were doing. And then in 2011, IBM Watson won on Jeopardy. And I was curious, like what that was and how it worked. And I was trying to solve marketing strategy and budget allocation at the time. And, and so I looked at Watson and thought, I wonder if that could help me predict what to do. So when a company would come to us and say, we want to grow leads in this industry, how much money is it going to take? I was at the point in 2011, 2012, I was like, I don't know, there's like 10,000 ways to spend your marketing dollars. Like when I came out of college in 2000, there was 10. So I initially saw AI as a possible path to build more predictive models around what marketing strategies should be. And in the process of trying to figure out where AI actually was and what it was capable of doing and what it even was like to be able to have a reasonable definition of it, I became extremely curious and passionate about where it was going to take business and society. And so my second book in 2014, I wrote like 700 words out of 50,000 about artificial intelligence and some theories on like this idea of an intelligence engine that would make these predictions about outcomes and behaviors. And that became the only thing anybody wanted me to talk about. So the rest of the book was like, it was all talent, tech and strategy for marketing and the AI thing. I started getting calls from all these companies. The US government wanted to talk about AI and recruiting for the military. Like it just sort of took a life of its own. And, and so in 2016, basically we wanted to keep researching it. And so we created Marketing Institute as like a, a separate component of the agency. It was like a spin off of the agency in essence. And it was just a place for us to talk to these AI researchers and entrepreneurs and share what we were learning, like tell the story of AI. And that led to 2019. I split it off its own company, funded it, eventually raised a million dollar seed round for it, launched the conference, survived the pandemic, and here we are.

Well, and you Know, I don't want to go through the whole story in three minutes, but I do remember probably, I think 2016 or 17, you were doing a workshop on AI social media marketing world and hardly anybody was interested in it because it was, it was kind of like little niche thing. Right. And you know, you started this conference. Right. And, and people, this is a great story of perseverance. Right. Because there were some times there where I would imagine you were a little early. Did you feel early and kind of, how did you survive that? And then ultimately what was the cat? I think I know what the catalyst was, but what was the catalyst that kind of blew it up for you?

Yeah, so it was definitely difficult years. You know, we lost money every month for almost eight years, or six years, I guess seven years. So it was definitely a long play on my part. So in 2021, April of 21, so it's a year and a half before ChatGPT, we were on vacation and I was listening to Genius Makers by cademetz. So it was a book that told the story of like modern AI from basically 2011 until 2021. And I knew a lot of it, but I. There was some pieces of the story that I just hadn't heard before. It was the first time that the story was ever being told because CADE was deeply sourced with like the AI leaders of the time. And so I came back from a walk when I was listening to it and I told my wife, I have to sell my agency. And she said, why would you sell the agency? Said I overestimated how quickly AI was going to be prevalent in the world, but I underestimated the total impact. And I have to spend like every waking second figuring this story out and not wasting my time on other people's businesses. And so I decided like, that was it. I was done. I was gonna, you know, go now. We were still losing a bunch of money at the time, but I had raised a million dollar seed round by that point, so I had, I had time to burn. But we were in the middle of the pandemic and things weren't going great. We didn't even, we couldn't run our conference the year prior. 21 was a virtual, you know, event. You know, it. The event world was not, oh yeah, the place to be.

Those years, those were rough times.

Yeah, yeah. So we survived. And then when Chad GPT hit November 3rd of 2022, I had about four months of burn rate left. So I was going to be out of money by February, March of 23, Chad CPT hit, and things just started taking off right away. Like, I was starting to do phone calls from executives at major companies, heads of VC firms, presidents of universities. Like, everyone within the first two weeks after ChatGPT started calling me, saying, okay, that stuff you've been talking about the last, like, six years, what is it again? Like, what is it? What is going to happen? And so we started getting, like, qualitative feedback that it was kind of taken off. And by January, February, we started seeing the quantitative follow, like site traffic, podcast downloads, leads, people signing up. Our Intro to AI class, we teach every few weeks. We used to have two to 300. We were. We were getting 900 to a thousand all of a sudden, every other. Every third week. So I had a pretty decent indication it was going to take off. And then I think March of 23. So I founded the institute in summer 16. March of 23 was the first month we ever made a profit. And then it just sort of took off from there. So, yeah, I definitely, you know, barely survived. And I think it was a combination of conviction. Like, I. I believed this was going to happen. I just didn't know when I was going to mortgage everything I had to keep us there. Like, even when I was. I was four months out. I still have my whiteboard to my right, so that has like a. I can take money from here, be here, see here, D, here, all the way down. If I didn't raise more. So we were going to raise a big, big slug of money, or I was just going to fund it myself one way or the other. We weren't closing the institute when I ran out of money. We were just going to find more.

So bring us up to the present. Like, what's happened since the Chatbots moment? I mean, like, now, where is the business?

So our conference, we had 200 people attend in 2022, the first year we were back in person. And then 2023, we had 700. And then this year we're expecting 1500-2000 maybe. You know, it's hard. It's so hard to predict at this point, as you know, I will be there.

I will be there. Everybody who's listening, I plan to be there. I mean, that's amazing. And I would imagine that's. Is that the primary source? I mean, I mean, you don't have to disclose a lot of it, but you're offering other services as well, right? Or is it just the conference?

Yeah, I think of our business as it's currently constructed as a media event and education business. So the media side is podcast newsletter downloads. We about 75,000 subscribers. And then the event business, we have our main conference and then we have three virtual summits right now and then online courses. So those are the main things. And then speaking and consulting and workshops make up another chunk of revenue. That was not really a thing before 2023 and that exploded last year too. I mean I did price 70 keynotes in 2023.

Well, first of all, congratulations on persevering because the truth of the matter is that folks, Paul has been at this a lot longer than just about everybody else that you're going to ever hear on this POD podcast. He has been dedicated to this, educating and doing the kinds of things that, that people like me have been trying to do for a very long time in the social marketing world. And Paul, you got nothing but green fields in front of you. So congratulations on your perseverance. I want to now ask this question, a why question. There are marketers and creators and entrepreneurs that are listening to today's podcast and maybe they don't fully understand why they ought to pay attention to AI and maybe they don't kind of see where it's going. What do you want to say to them?

Yeah, the way I think about work knowledge, work in particular, so you know, I don't, I know you have a global audience. I'll talk specifically about the US economy because I know it best. There's roughly 136 million full time workers, full time jobs in the United States. About 100 million of those, give or take, are knowledge workers. People who think and create for a living. Could be engineers, could be attorneys, doctors, freelance writers, graphic designers, marketers, all of us. I am under the working assumption that at least 80% of what those people do will be AI assisted to some degree in the next one to two years. Now what that means is the software we use to do our job. I mean in recent time we saw Adobe is infusing AI into everything they do from Firefly to the integration of video generation models. We have obviously a bunch of large language models through anthropic and OpenAI and Google and Amazon. Like everyone is creating all of these technologies and those technologies are going to be infused into everything you do. Your social media platforms, analytics platforms, CRM platforms. It's going to be impossible to do our jobs where AI isn't there helping us. Now to what degree it helps us and how much of our current job the AI does for us is kind of the big question we all have. So you can go and look at your job responsibility day. And let's say there's 20 things you do every month. You have these standard reports. You create, you write things, you design things. Like, whatever those things are, you can look at those and say, okay, how much can I help me right now with these things? And it may be 10 to 20% of the work, which is a huge lift. Like, that's a lot of help. A year from now, it may be 50%. And so this is this dynamic environment we're in where AI is going to be able to enhance what we're capable of doing, but it's also going to automate, like, truly automate some things that we just won't have to do anymore. And so the faster you kind of accept that this is the world we are in and the world we're heading into and get over any fear and anxiety you may have about this, because it's not going to go away. It's only going to get smarter and better. Then you can actually start to figure out how to build a more fulfilling career because of this technology and a smarter company, a more profitable company, a company that does better things for its employees. And so the only way, though, to do that is to, like, get into the tech and see what it's capable of, apply it to the use cases, the tasks that you're doing every day, and see the ways it can help you. Like, a podcast is a great example. It's one of the ways at the Institute that we very aggressively infuse AI, because our podcast is our primary audience growth engine right now. And so if there's 25 things we do each week to plan and produce and promote the podcast, AI is helping with, like, at least half of those probably saves our team 20 to 30 hours a week. And honestly, like, the 20 to 30 hours is things we may have just not done if we didn't have AI, didn't replace anybody's job. An example would be take the hour video and use an AI tool to cut up 10 shorts that can then be published to TikTok and YouTube. Like, that's just something we probably wouldn't be doing right now that we can use to accelerate our audience growth.

Looking into the future, a couple of years from now, and you're just talking about, like, the near future, right? Where do you see AI helping? Especially people that adopt it early, five years down the road, do you feel like there'd be an unfair advantage to those that begin to learn and adopt and experiment it with it now?

Yeah. Five years in the in the AI world is so hard to like, that's.

Like decades in the regular world it is, yeah. Okay, three years.

Yeah, I think, like, I generally speak in, in terms of months, and then like one to three years I think is maybe reasonable. Like, I mean, I spend every waking second thinking about this topic, listening to podcasts, reading books and research reports like, and I, I think I have a reasonable estimated, like, educated guess on what the next 18 months could look like. I think that anyone who used or was aware of GPT3 and what it's capable of and what its limitations were, and then what we got when GPT4 showed up into the world in March of 2020, 23, all indications are that the next levels are going to keep taking leaps as big, if not bigger in terms of how smart these models become, how good they are at thinking, at reasoning, at understanding. They're going to become true. Like strategic assistance that you can rely on to help you build plans and analyze data and insights. There's going to be a massive reduction in error rate. So when you use it to help you create things, it's going to be highly factual, probably more factual than the average human is. They're going to fix all of the deficiencies with AI today. So what I always say is like, this is the dumbest form of AI we'll ever have in human history. It only gets better. And so a lot of those improvements that I just said are likely on the slate for the next six to 12 months. Like, all the next versions of the biggest models in the world are going to make a lot of progress towards solving what we see as the current deficiencies. Then when we move into 2025, 2026, you're going to start hearing a lot more about this, your AI agents. So not only things that can write things for you and help you plan and ideate, but things that can do things for you can go take actions. Google recently announced their Vertex AI agent builder, where you're going to be able to go in and just build your own agents, much like you'll be able to do an OpenAI and others. And so this idea that you and I and I have no coding ability, I don't know about you, Michael, but like, who have no coding ability whatsoever can actually build the things that are in our heads. And so I just think that the sooner you get in and become competent in these technologies and comfortable with them and comfortable with the change they're going to keep bringing, it just continues to amplify your ability to stay ahead of your peers and become insanely valuable within your company.

Awesome. Okay, let's assume that everyone who's listening has bought in on the importance of AI, but maybe they're just about to start the adoption curve. What are some of the things we need to be thinking about and the things we need to be considering before we adopt AI? And let's focus mostly on the smaller. Maybe they're solopreneurs or maybe they're small businesses. What are the things they got to be processing?

Yeah, first for everybody is education and training. And then that. That just keeps going across your learning and adoption journey. But listening to a podcast is great. It's a kind of a Kickstarter, maybe for you, or the first time you hear someone give a talk or the first book you read, or whatever it is that gets you kind of going and gets your curiosity going. You got to stay on that path. And everybody always asks, like, what's the magical formula to learn this? I don't know. Like, the magical formula to me is, how do you learn best? Is it podcasts? Is it books? Is it online courses? Is it a mix of those things? Whatever it is, take the time to highly curate those sources and then just get in there and learn. And then pick one or two or three tools that you are going to go deep on. Not put five prompts into ChatGPT and be like, I don't get what all the buzz is about. I'm not using this. No, like, pay the 20 bucks a month or whatever it is for a ChatGPT team or individual license and grind on it. Like, really give it hard. Use cases like, say you're planning a conference. Have it write abstracts for you. Have it help you come up with an agenda. Have it help you identify speakers. Like, do the things you're already going to do, but have the AI work hand in hand with you to do these things. And you will start to see ways to blend the technology into your workflows and that. I don't care what size business you're in, that is like, the main thing you need to do is comprehension is like understanding. And then competency comes from experimentation yourself. You cannot just listen to other people talk about what they're doing and their tasks.

Let's talk a little bit about a concept of, like, preparing our minds, if you will, before we dig into this thing. Right. Because I have experienced this and I'm sure you have as well, where people are like, oh, I've tried ChatGPT. It was the free version. Right. It's not very Good. How ought we like prepare ourselves before we actually start messing around with this tool? Is there any kind of like reprogramming of our presuppositions or our expectations before we start experimenting? I'm just curious what your thoughts are on that.

Yeah, I think people who have that experience like you described, where it's just like, yeah, I don't know, wasn't life changing for me. They probably haven't properly prepared themselves. I think as you're illuminating for what its true capabilities are and what the difference is between the free version and the paid version, you know, what the capabilities are. So I think some very basic concept of like, how do I prompt this thing? Honestly, like you can go take a bunch of like online courses or read a bunch of stuff. I found the simplest way to explain it to people who've never done it. Talk to it like a person. Like, right as though you're giving a project brief to an associate or a smart intern or something. Just tell them what you want it to do. Give it examples of what you're trying to do. So I want to write an email promoting my upcoming event, my upcoming webinar, whatever it is, my product launch. I'd like some help doing this. Here's examples of the five best performing emails we've previously sent. Can you write this in a professional tone? Just tell it as though you're handing it off to another person to do and you're expecting a really high quality first draft from. And then when it comes back, talk to it like a person again saying, hey, can you actually make this a little longer? Like expand on this section and make sure to again, just tell it like you would like manage the AI. And I think if you just do that and ignore all the other guidance about what to do in prompting and which model use. If you just do that, you will get results. Like you will get valuable output from these things. And I think some people just don't even take those steps. Like they just get in there and it's like, you know, back in the early 2000s, like I don't know how to search in Google for anything. And back then you had to be pretty decent at searching to like get something out of Google. I think that's kind of where we're at now is like people get there and they look at it like I don't know how to prompt this thing. And they do try one or two and it's like, ah, that didn't work. I'm just going to keep doing my own thing.

I've got some thoughts on this. You know, you and I are both have backgrounds as writers. You have a journalism background. I don't have a journalism background, but I wrote a book on writing. If you have a writer's background, you have an unfair advantage because you understand how to query people and interview people. And I would say something like this. If you were sitting in front of Einstein and he was still alive today, and you asked him a very stupid question, a very simple question, you're probably going to get a very simple answer, right? But if you take your time to think about how to ask the question to get what you want out of it, I think you're going to increase the likelihood. And you know, most people do what I do. They say, who is Paul Raitzer? Or they say, who is Michael Stelzner.

Right.

But where it gets, I think, really, really intriguing is asking it, what else can you do? Do you recommend that? Because I was messing with Claude and I uploaded a closed captioning file from one of my podcast interviews and I said, write me an article from this closed captioning file that's 1500-1700 words long. I was kind of blown away. But then I started asking it more questions. Where would you suggest I insert images? And it had strategic locations where it said, well, in the interview when you mentioned this, I would put this image here. And then I went on to ask it, what are the tools? And it mentioned the tools and what are the URLs of the tools? And it mentioned those. And then I said, what else can you do? And it just said all sorts of stuff. So it's hard when you're just looking at a flashing cursor, right, Paul? To even know what to ask. I mean, these things are much more capable than we imagine. And if we were talking to a human, they would ask us questions and we'd go back and forth. But that's not generally how these things work, Right?

Yeah, I agree. And I think, for example, like, if you're not sure what to do and you go go in and say, I'm working on an email campaign, I need to write a three part emailer that drives them to sign up for this event. What do you need to know from me to do this? Like, yeah, like, I think if you just, to your point, like, have this conversation, let it tell you what it's capable of doing. It may say, yeah, give me some examples of like, what you're looking for. Okay, great, now give me. Or it may say, hey, let me take a first draft and See, see what you think. And then I'll say, what do you think of this? So yeah, I think there is this whole idea of like, you have to just get in an experiment because the models work roughly the same. Like, it's kind of weird that they all sort of work this way. But there was a recent Google published a prompting guide recently for Gemini, which is their big model, and they found that the average prompt, the best performing prompts, were 21 words. So like, up until I read that, the expectation was, no, no, no. Like, give it all this detail, give it a bunch of examples, give it all this stuff. And what's happening is these models are getting so smart. And in essence, what happens behind the scenes is they're able to improve whatever prompt you give them. And they may not tell you they did that, but basically it's like rewriting the prompts behind the scenes so that it optimizes the output. You can see this at work. If you go into ChatGPT and have it create an image for you. You say, I'd like an image of a safari and a robot walking through or something like that. And it creates it. It'll say, here's your image of a robot walking through a safari. If you say, what prompt did you use? It will actually show you the prompt it rewrote from your weak human prompt. And it'll be like a hundred words and have all this detail, these adjectives and everything, and you realize like, oh my gosh, these things are like superhuman at prompting. And we don't even know that it's happening.

Quick question. How many podcast episodes have you recorded? How many webinars have you hosted? How many client calls or team meetings have happened that are now just sitting there in your files? Here's the big unlock. Those aren't just old recordings. They're full of insights about your audience, your brand voice and your customers actual objections. You just need to know how to extract them. That's exactly what AI educator Erica Stanley is teaching AI Business Society members on January 8th. By the end of this training, you'll have a repeatable workflow that turns every past recording into a content and data powerhouse. One that keeps paying dividends long after the mic's been turned off. This training is exclusive to AI Business Society members. But here's the good news. You can join us right now by going to socialmediaexaminer.com AI and start transforming your old content into new marketing gold. SocialMediaExaminer.com AI to join right now. Most people, when we think of Generative AI are thinking about the production of text. You know, you're asking it to write something for you. Let's explore some maybe not so obvious use cases. They're probably obvious to you, Paul, because you study this so deeply. But what are some not so obvious use cases that we ought to be thinking about? Maybe let's just start with data analysis. I mean, like, can this thing do data analysis?

Yeah, because these models, I don't know if all of them currently do, but I, like, I know Gemini does, I know ChatGPT does. I think Anthropic does. They all have vision capabilities so they can see things. So you can take a screenshot of a podcast chart or an analytics chart and you can give it to it and say, analyze what's happening here and it'll do it. Wow. Okay, so that's a cool one. And it's fun to kind of mess around with. And it's kind of almost shocking sometimes how good it is. So, like, I did this for a podcast example, and we had this dip in December, and so it looked at it and it's like, hey, what's going on? Like, something happened in December. And I was like, well, what do you think that could have been? And it came back and said, well, you know, the holidays are over December, so it's a chance that it might have just been a holiday flux. And I was like, that's exactly what it was. And the same thing happened in July, and it said, actually, it looks like that happened in July. So, yeah, I think just using it for vision capabilities, I've demonstrated before in courses and in talks where you can give it samples of ads or social shares or your landing page or your webpage and ask it to analyze it. And then if you give it context, if you say, like, hey, here's the 10 criteria we use to assess landing pages. Can you look at this page and tell me how it's doing? It'll do that, and then it'll give you specific recommendations for, like, CTA and colors and positioning of images and things like that. So again, it's that just base understanding that, oh, they can see things and understand what they're seeing. That's a really interesting one from a vision perspective.

Yeah, We've successfully had it give us feedback for split testing on a sales page that we were doing. Like, it suggested that it was a little bit too long or maybe we ought to include pictures with our testimonials. So is it true today that generally you can give a link, or do you mostly have to generate PDFs and upload them or what's the. And again, I know this is all going to be potentially dating this part of the interview, but what's your thoughts on the way they use today? Do the most of them do external links and can scrub the data.

Yeah. So my current understands CLAUDE is not connected to the Internet. So anthropic Claude is not. Gemini is. ChatGPT is. So you, you can give them URLs, screenshots. Screenshots are fine. PDFs fine. One example I thought was as you were saying that was like, let's say you wanted to do a pitch deck or even like a product deck or something, or a media kit to sell sponsorships. You couldn't there give the PDF there and say, hey, listen, our goal is to sell sponsorships to this audience. Can you analyze this presentation and tell us how we can strengthen it and make sure you consider text visuals so you could do this for any presentation you're giving, probably upload. I haven't tried that, but actually I may go do that because that would be a really interesting one.

I've also had my sales team use it for objection handling. So they will literally put in the product they're selling the objection they had from their prospect that they received over email, and then they ask it specifically for different ways to respond. I've also heard people can use it to kind of make emails better. I mean, I don't know, do you do this? I mean, you and I are writers, so we don't think this through. But is it pretty good at suggesting edits for basic copy?

I don't use it for that, but my team does. Yeah, I mean, you can take your drafts, you can take like, you know, for us, like what we'll do, what we used to do for our events was, you know, you start, let's say you're starting to promote the 2024 event. You go back and look at 2023 and 2022. You spit out a database of all the emails that were sent. You look at the best open rates, the best conversion rates. Like it's this heavily manual process to like, okay, what emails worked before? Was it featuring speakers? Was it featuring offers? Was it basically just the time of like a price increase was coming? Like, what was the thing that did it? You could take that database in a CSV file and just give it to ChatGPT and say, hey, we're trying to figure out what emails work best. Can you first analyze this? Okay, great. Here's what it's finding. All right, here's the examples of those emails that work best. Can you draft us a sequence for this year based on this or here's what we already wrote. Can you tell us how to improve these? It absolutely can do that.

There are some considerations that people need to keep in mind when they're using AI. Let's talk about a couple of the obvious pitfalls. Let's start with company confidential information. What's your thoughts on that?

Yeah, so that's a mistake a lot of people make. I think up front is especially using like a free version of something and you don't have a terms of agreement where you know your data is staying your data. Some people have gotten in trouble in the early days of Genai by taking a transcript from a Zoom meeting or whatever it is where there was confidential proprietary information talked about. Drop that transcript into a chat GPT and ask it to summarize it. Well, that confidential data now lives in OpenAI server somewhere. They may or may not ever use it to train anything on. But the reality is you've just basically leaked confidential information. So you have to be aware of the tools you're using and what the terms of agreement you have with them are. This actually, now that I'm like kind of saying this out loud comes into mind too when people have their like note takers show up for meetings that you haven't invited.

Yes.

And then you realize like, well, I don't even know what your note taker tech is. And now anything I say is going to end up on the servers of some third party company. So there may be big enterprises that's a no, no. Like they're not allowed to have conversations with some random note taker on them. But it's kind of like, I don't know if you remember this, like before the pandemic, you were never sure when you got on Zoom if videos were going to be on or off. It was kind of that like, are we doing videos or aren't we doing videos? I'll let, I'll let Michael go first and if he turns it on, I'll turn mine on. Now it's just like, Zoom is video. I feel like we're kind of at that stage with our AI note takers right now where it's like, is it cool if I turn on the AI companion in Zoom? Like, is Michael going to care if I'm recording this or do I have to ask Michael if I'm recording this? In the future, I think we'll just assume every meeting we're having is being recorded by something, but I don't think we're there yet. But that is definitely a consideration. Is your data proprietary, confidential information? Make sure you understand terms and that you're protecting that stuff. And I always say talk to your attorneys is like my main thing.

What I've been able to do when I'm doing data analysis, like I've done a lot of data analysis inside of Claude by Anthropic and I'm uploading CSV files and I'm anonymizing the information, right. So I'm taking out personally identifiable information, names, email addresses, phone numbers. Because that stuff, you don't want that out there, Right. And you've got terms of service privacy stuff probably in place for your company. So that's something important we need to be thinking about. I've also told my company that they're not allowed to put any confidential information into the AI systems unless they're using the API, which is that still true? I mean, do the paid versions of these software applications generally not put the data into the large language model, or is it only if you're going through.

The API, supposedly it's protected?

Is it okay?

Yeah, because they knew they were going to get adoption in enterprises unless they did do that, so.

Okay, well, that's good to know. Let's talk about proper use of output that comes out of these systems and some of the pitfalls people might not be aware of, like, for example, copyright in America. Talk to me a little bit about that.

So the main takeaway there is, as law currently states, only a human can own a copyright. And a prompt of an AI system is not authorship. If you use ChatGPT or anthropic, Claude or Gemini to write something, say a report or a blog post, you don't own it. You don't own the copyright to it. Nobody does. There's always the question, well, who owns it? It's like, no, nobody owns it. So you can commercialize it. You can use AI to create things again, research reports, articles, blog posts, whatever you want to create scripts for things. You can create it and you can commercialize it, but you cannot protect it through a copyright with the U.S. copyright Office. Now that may evolve. They have been going through listening sessions for the last year. I know that they're considering adapting it in different ways, but I don't know of anything that's changing anytime soon on that.

Yeah, and this is where it gets really tricky, right? I like the fact that you can commercialize it. That means you can use AI to assist you in the creation of a course, for example. Right. And you can use AI potentially in the creation of a paid product. But the moment where you cross into some fuzzy lines is if you're using AI to write a book, you know, some people might be wondering, right, here's how I think about this, and I want to know what your thoughts are on this. We've been using tools to alter images forever, right? And you take a picture, you use Photoshop, you alter the pict. And I would imagine there's gotta be. And neither of us are attorneys, but I would imagine if you're using AI to help you come up with an outline or to help you do some research. Right. And just to get started, that's probably less of an issue than having AI create an entire piece of work for you. I don't know what your thoughts are on that. I'm curious.

Yeah, for sure. No ideation, summarization, transcription of your words. Like all those things are, in my opinion, and from having talked to a number of attorneys, those all are good use cases for AI.

So, for example, this interview, if we use AI to create an article because the source material is us, that feels more legitimate than asking the AI to create something from the ether. Is that what you're saying?

Yeah, and we do it all the time at the institute. So we take the transcript, we use the AI to. To not only create the transcript, we use AI to summarize the transcript and write summary blog posts of it. We then use it to create social media shares from the transcript. So for us, the original authorship is the words of the podcast, and then from there we're just using technology to propagate those words out to different audiences. So I view that as a totally safe and viable way to apply AI. Now, if we got to the point where we were scripting the podcast completely and I was using AI, so in our case, Mike interviews me on our podcast about the main topics and rapid fire items in AI each week. And so if he asked me about Google's Gemini model and I just went in chat, he said, write me a summary of it. And I just sat there and like, read it verbatim. We got a problem like that. That is not authorship.

I seriously. Yeah, I know one of the things you talk about a lot is setting some ground rules for your business when it comes to AI. Can you give us some general advice on. Especially for those of us that have. It's more than just ourselves. What are a couple things we need to be considering when we deploy if you will, AI inside the workforce.

The most important one that I would focus on, whether you're a one person shop or thousand people, doesn't matter, is generative AI policies. And it can guide you individually or can guide your whole team. And that is basically like in a nutshell, which tools are we allowed to use? What use cases are we allowed to use them for? Do we have to disclose that we have used them? So if I'm using AI to help me reply to emails, do I have to tell the recipient I used AI to help me reply to an email? Andrew Davis did a great talk at your conference about this and the Drudini thing where he created like his AI avatar in essence, and he was using it to reply to emails. So the question I was like, well, do you disclose that when I go to your website, if you're using AI on your blogs, do you say, hey, this was 20% written by human, 80% AI, or do you just have generative AI policy that just says, hey, we use it to help outline blog posts, we use it to whatever. So those policies are just, they're not meant to restrict people from using AI at all. There's actually meant to give them the freedom to use them with the basic guardrails that you have decided are going to keep your data, your privacy, the secure information. So you put that stuff in place, you bounce it off your attorneys, make sure everything's good to go, that you're using it properly, using it within laws and regulations, and then your team now is like, okay, cool, like I'm allowed to use it to do summarization, transcription, ideation. If I use it to help with blog posts, I'm supposed to use it in this way, not that way. We are not allowed to use AI generated images on our site. We are not allowed to use AI generated video in our promotions. Those are all the things. I think there's basically like 10, 12 things you got to kind of walk through and say, yes, no, yes, no, yes, no. Here you go, team. This is our current policies around this stuff.

I got a couple of questions about something you just said. You said, show it to your attorney. Now, is this because we're going to publish this out on the web, or is this because you just want to make sure you set yourself up for success internally in the company or both? I'm curious what your thoughts are on that.

A lot of times, at least in my career, marketers viewed attorneys at a bit of a roadblock and like slowing things down. The reality here is there's some major pitfalls if you do this wrong. So if you scale up your blog just using AI and you get 50 posts in after six months, and then your attorney's like, we don't own the copyright to any of this, or you're a brand and you're outsourcing this to like a freelancer or an agency, and unbeknownst to you, they're using AI to create everything that they're doing for you. And now you don't own the copyright to anything they gave you or trademarks on designs, things like that. So you really just need to know your own policies, but then the policies of the people you work with and to make sure everyone's kind of working under the same umbrella we just talked recently about. There was one big agency said they'd won three big deals recently, and every one of the master agreements said that they weren't allowed to use Generative AI. So we're entering this phase where there's like this uncertainty around are we or are we not? Are we telling people we're doing this or not telling them? But it's really important that it's one, you're on the right legal grounds that the stuff you create you own if it's important to you to own it. And two, that the people you're working with are working within the policies you've set for your company as well.

What kind of an attorney do we even hire if we were to hire an attorney? An IP attorney. Okay, yeah. And then do you or do you not recommend publicly disclosing AI and some sort of policy on your website?

I do. I actually think it's going to be like a terms of use in the future. Like, every site should just have a Generative AI policy link on their site. And you go there and it says, hey, here's what we view Generative AI as X, Y and Z. We use it to do this, this, and this. We do not use it for that, that and that. Give me super simple. But now it's just like super clear. And if someone's like, hey, did you guys use AI to create this? This video looks like it was AI. It's like, no, here's our policy. Boom. And it's not like every single post you do or every piece of content you share, you have to, like, disclose how you did it or didn't do it. You just have a policy that states it.

So is that policy that you put on your website the same policy that you're using internally? Or is it more like modified? More for consumer consumption, I think it'd be modified.

Most likely you may have. It's going to be real similar, like, this is how we use it and don't use it. But your people on your website don't care if you're using it for internal emails or not, or internal proposals or like things like that. So it's probably going to vary slightly for the different audiences.

When you think about yourself, Paul, how much time each day do you find yourself using AI tools? I'm just curious. And, and then also how are you using them? Because I think that you are probably more attuned to all the possibilities. And on one side I can imagine you want to avoid it because it's a rabbit hole that you could go down. On the other side, you might have certain kinds of tasks that you're used to using it for. I'm just curious, like, how is AI helping Paul be a better CEO and creator?

Honestly, probably less than people would assume. But I also don't do a bunch of data driven, repetitive things. Like, as a CEO, my days don't look very similar. So, like right now I'm in production of a new course series. My life's going to look similar for four weeks. Like I'm going to be doing very similar things and I will be like, very much integrating different language models to do research and to develop outlines for things and to ideate. So that's infused into that course creation process. When I'm planning agendas, I'll use it to help with abstracts for sessions and things like that. So I have it when I'm in doing specific things. But a lot of my daily life is more consuming AI information in a lot of different modalities, from research reports to YouTube videos, to podcast interviews to whatever it is and synthesizing that information for our podcast and for my own business and for the limited consulting work we do. The only way I really use AI in that process is when I hear something like, there's something I was just listening today about the law of thermodynamics. I have studied the law of thermodynamics like 10 different times and I couldn't. If you ask me right now to explain to you, I can't do it. Every time it comes up in relation to the future energy consumption needed to power the chips that make the AI work, I have to go back into ChatGPT and say, give me the law of thermodynamics again. Like, I'm a seventh grader. Like, that's how I use it, is to like understand concepts so I can relate them to other people. And so I always have to try and like take these complex ideas and boil them down. So I use AI all the time for that anytime I'm doing brainstorming strategy always. And I'll usually use all three different core models because I'm always like seeing which one's better than the other at this right now I don't really use it much for writing. I don't use it in my emails. My team uses it dozens of times a day for everything they're doing. But I don't personally use it a ton other than kind of those use cases I talked about and then a lot of just experimenting with new technology to see what it does.

Paul, you're a really smart guy and you're a deep thinker and I think of the Gartner hype curve. I don't know if you're familiar with the Gartner hype curve. Where do you feel like we are on the hype curve here with AI right now? Is this just. Are we in a massive hype cycle here and we're about to enter the trough of disillusionment or what's your thoughts on this?

So I've seen some people whose opinions I respect say we're in the trough of disillusionment and I would respectfully disagree with them. I definitely think there are people who are just like, oh my God, like I have a fatigue, it's, it's overload. I can't deal with this. I already got a full time job. I don't want to have to learn AI too. And the way I kind of address that is like, what would you say if it was in 1994 and the Internet was coming? Would you say I don't have time for the Internet and like let someone else figure the Internet out? Like, right. This is going to transform our lives in ways well beyond what the Internet did. We don't get the choice to not care and to not like dive into it. So I actually feel like however you would think about it, within Gartner's cycle, I think we're so early and so under hyped in terms of what it's actually going to do. Because enterprise adoption is just starting. Like we just middle of last year started getting enterprise platforms to do these language models on and even now it's still early in that adoption cycle. I think it's going to accelerate really fast. But most enterprises haven't really figured this out yet. They're still piloting different stuff, they're not all in on you, even the ones that, I mean, I've talked to organizations that have thousands of copilot licenses with Microsoft and they haven't even rolled them out yet. No change management plan tied to them. It's just they went and got the tech because it was there and it seemed cool and they would figure it out. So I just think we're still early and I think most business leaders don't comprehend how transformational it's going to be and how fast it's going to happen.

Well, and I think about, I started Social Media examiner in 2009 and I was late because we already had LinkedIn, Twitter and Facebook and they've been out between three to six years. We're not even a year and a half as of this recording into the launch of ChatGPT as we know it in the world. So it's very exciting to hear that we're early days, but it's also very, I think, sobering for everyone to understand that this is faster than anything we've ever experienced in our lifetime. And it's really, really important that people learn. And Paul, I definitely am going to advocate for people to listen to your podcast. So tell them where to find the artificial intelligence show and also tell them where they can find Paul Raitzer if they want to connect with you on the socials or if they want to reach out, learn more about your company.

Yeah, marketing AI institute.com is the home base for everything. So you can go to resources. The podcast is right there. You can find that all the podcast networks and it's also on, we publish on YouTube as well. So you can watch the episodes if you prefer. Courses, newsletter, events. It's all on marketing AI institute.com and then for me personally, I'm most active on Twitter, slash X and LinkedIn. I'm not personally very active on Instagram. I think I actually have a private Instagram Instagram account and Facebook. I I go in when I get birthday wishes, basically.

Paul, thank you so much for coming on the show today and helping open our minds.

Thank you, Michael. Yeah, love it. Thank you.

Hey, if you missed anything, we took all the notes for you over@social mediaexaminer.com a1 Be sure to follow this show on your favorite podcast app. And would you share this show with a friend? You can tag me. I'm at Stelzner on Facebook and ikestelsner on Twitter X and do check out our other shows, the social media Marketing podcast, the Marketing agency show and the social media marketing talk show. This brings us to the end of the AI Explored podcast. I'm your host. Michael Stelzner will be back with you next week. I hope you make the best out of your day and may AI help you become more successful. The AI Explanation Explored podcast is a.

Production of social media examiner.

AI Business World, early bird. Pricing ends on January 16th. You can save $300 on your tickets. Register at AIbusinessworld live today.