15 | The AI Advantage: Business Strategy for Winning in the AI-Driven World, a Live Episode with Paul Roetzer of Marketing AI Institute

25:21But the way I'm going to do this and the way I'm advising other people to do this is I would take their job description. So in this case, we're taking the job description of a future hire, but you could do this with the existing. and I would go line by line of the things that they do. So tactically, what are they doing? If they're in charge of sending the email and newsletter each week, writing and sending the emails and newsletter, they're in charge of crafting social media shares based on the podcast, they're in charge of whatever. Whatever it is they do, and I would go through each of those things and say 12 months from now, how much of this is the AI doing?

25:54So again, this takes a little bit of a level of experience and expertise, understanding what AI is capable of to do this effectively. but you have to find a way to get to that state, either internally through outside advisors to where you can do this because you owe it to your team to do this. It's the only fair way to be clear about this. And I would involve them in the process at some point, have them do this, because what you want to do is say, okay, there's a chance a year from now, this email marketer's job is going to be 50 % AI. Like we can get this tool and they're going to be going from 130 hours a month of doing this thing to 65 hours.

26:32And that's just like straight up, the tech already exists. There's no reason to do this other than we don't want to take this person 65 hours. So what I would do is look at it and say, okay, we're a pretty clear path to reduce this person's full-time equivalent down to the 65 hours. What can we redistribute that time to? What is on our wishlist right now that we're not doing? Do we want to launch a second podcast? Do we want to launch a second email newsletter? How can we redistribute this person's time and experience to the benefit of them and the company versus just saying, oh, cool, there's half of an FTE we can get rid of?

27:05That's what I think needs to happen. And I think the more in the open it happens, the more involved people can be in raising that and saying, you know what, I'd love to do. What I see as an email marketer, I don't ever get the chance to do, but it would be awesome if we could, would be this and this. And so, yeah, awesome. I don't really actually enjoy the segmentation part. I don't like looking at the performance reports. I don't like doing this. Oh, great. AI can do those things. Let's have AI do those and let's have you shift over here. So I think you need to do that. You can think about it at an individual level.

27:34It's much easier to process that's a more tangible thing. You can do it at a team level. It's like a product marketing, let's say email marketing, advertising, events, whatever it is. And then you can look globally. Now you can look by a whole business function. And so I think individual managers, like managers, directors should do their teams, VPs, C-suite, they're looking whole business functions. And I would make that a real priority. I really don't think people realize how much knowledge work is going to be intelligent automated in the next 12 to 18 months. And my unscientific estimation, this is based on my own assessment of our team and some other roles that I've looked at is most knowledge work, you could see 20 to 30 % gains in efficiency within the next 12 to 18 months.

28:18That means if someone works 150 hours a month, you could see 45 hours roughly shaved off of that work by intelligent automation. Now do that across an entire team. You're talking about thousands of hours potentially. Nobody's ready for that. Nobody knows what that means. Nobody knows what to do about that. If that sounds daunting to you, welcome to the club. This is all, people are all figuring this out at the same time. No one is even really talking about this. Nobody wants to talk about the impact on knowledge work, but it's real. I agree. I want to touch on two points that are directly related to that.

28:53One is how critical it is to go back to the education and training part, right? Because your ability to do the process you just described has to come from a real understanding of what's possible right now and what might be possible in the next six to 12 months. Because otherwise you'll be making very bad assumptions and making very wrong decisions based on those bad assumptions. So the education part has to come first. And if you don't have that person, hire a full-time person to do that because that might be the most important person in the company or in defining the new trajectory of the company for the next foreseeable future, or hire a consultant.

29:35Bring somebody from the outside who knows that stuff for a period of time to help you through that process. So that's number one as far as I think. The other thing is, I think the scariest thing to me when I look at AI technology and where it's driving business and society, the bad part of it all comes from pure capitalism. But if you're saying, how can I maximize my profits tomorrow? A lot of people go, oh, I can do what I'm doing today and save, I'll go with your numbers, 30 % of the cost. And I really love the approach you take, which by the way, is the approach I really believe in. Let people thrive, involve the people saying, listen, this is going to happen, whether you want it or not, in the next, you said, 12 to 18 months.

30:19I think for companies who will do it quickly, it could be three to six weeks, especially if it's a smaller company and you can move fast. This is coming. I really want to keep you because I think you're the right fit for this team, for this company, and I want to support your personal growth. What other things you can do? Go research. Take two hours of your day every day in the next two weeks and look at AI tools and look at other stuff and tell me what you can do that you would enjoy doing that can help this company with AI tools that will complement the other stuff. And then instead of cutting cost, you're gaining new revenue opportunity and new efficiencies that you otherwise wouldn't have.

30:59So I really love these points that you brought up. There's a parallel I just thought of. Back in 2008, I had a big insurance company come to me. This is the very early days of social media and blogging and podcasting. All that was just rolling out in 2008, if people on the call remember those days. And they said, if we had a million dollars, what would we do? If we could get a million dollar increase in our marketing budget, and they didn't have a big marketing budget at the time. They had a very small marketing team, in fact. And they came to me and said, could you build us a plan of what we would do with a million dollars?

31:31So I did. And I almost see it in parallel now. If you go through and you do this assessment and you go, there is 10 ,000 hours a year that we think AI is going to do the work of our current team. I would do a plan of what could we do with 10 ,000 hours? So before I think, oh, cool, we can save 30 % of our costs, I would redistribute the 10 ,000 hours to say, what could we possibly come up with that could dramatically improve our culture, our team, our organization with those 10 ,000 hours? And I think if that was the focus of everybody is redistribution of hours to do more fulfilling things, more important things, then we're going to end up in a really good place.

32:10And I'm optimistic that there's going to be enough companies that take that approach, that this is a really real net positive for society, but it's not going to happen on its own. It really needs people that understand this and are pushing the same direction. Yeah, I agree. And before we jump to the last point, which is an AI roadmap, something else, I'm teaching a course to a bunch of CMOs starting in two weeks. And the premise of the course is really more about strategy, less about the tactics on what tools are available. And I think a lot of businesses, and I think that's why it ties up to the point very well, is you have to reevaluate your company's strategy.

32:48Why? Because the environment you're going to be operating in, the needs of your clients, the capabilities your clients will have in your house are going to change change dramatically. So the strategy you have today may not be relevant at all a year from now. And it's never happened in history before that something that dramatic happened that fast. So when you build an AI roadmap, and I'll let you, Paul, talk about this in a minute, one of the things you have to consider in addition to all the internal things we talked about, is my company strategy even relevant a year from now, a year and a half from now?

33:23And if it's not, what do we need to change? How do we pivot our services, our products in order to fit this new world that we're walking into? But now I will let you talk about what's in an AIO roadmap, how do you structure it, and who needs to write it, and what's the execution process of it? Just to build on what you were saying, I've actually had this conversation with some software companies recently where I was asked to talk to their product teams. And it's a really hard conversation because just as an example, a software product team, their job is to do incremental improvements that meet customer requirements or requests or popular items that the customer base has said they love.

34:04And you build within the confines of existing technology and you're really busy and you don't have time to be testing every new AI tool that comes out or keeping up with every AI announcement each day. I do it for a living and I have trouble keeping up with all that. So I get that there's this frustration within product developers because they could work on a feature for three months and some college kid will build it over the weekend and put it out for free. That's a real thing. This is happening to product teams daily right now. And so I think in that case, you really need this outside perspective and maybe just outside the product team.

34:42I'm not even saying outside of the organization, but you need someone else who's looking at what we call like first principles thinking. Does this feature even need to exist? Like we have five people working on this feature. If we think about where AI is taking us in five months, does this feature even, is it needed? And that's a hard thing. It goes back to education, goes back to understanding, to be able to look at things and say, I don't even know we should build it this way. And that's what I'm doing a lot of right now. Like with our institute, we're a media event and education company. I'm going to first principles on all three of those things.

35:14I don't want to build any of those the way companies have been built before. So how do we think about everything through the lens of AI? And let's just build a more intelligent organization from the ground up in those three core areas. So the way I think about it is you start with a vision statement. Okay, one year, three years, five years from now, this is what our organization looks like. This is what we're going to become. Here's how we're going to infuse AI. It's really helpful to do it at a user story level or like an employee story level where it's, okay, this person's job today looks like this in customer service and marketing and sales and product, whatever.

35:46In a year or three years from now, here's what their job could look like. So I would make it as real as possible, kind of like the whole idea of personas and market to one person. I would do that with your team. I would really build that vision down to the individual persona level of what each person's job looks like. You need to have a sense of what's going on in the core areas of innovation, like language, vision, predictive capabilities. So once you understand what AI is capable of where it's going, now you can better predict out what's going to happen. Then you get into your own organization.

36:17And does anyone on the team know this stuff? Do we have anything to build on top of? Have we done some pilot projects already that maybe we're not even aware of? So you take stock of what's going on, what's your current state of AI awareness and capability is, your staff, your talent, your tech. Then you go into what are the, we always talk about use cases and problem-based, like problems that we can solve. So use cases are things like, we know an AI writing tool would help. We're spending 500 hours a month writing. We don't need to spend three months to figure out an AI writing tool is a good starting point.

36:48So let's have a team go find an AI writing tool. And now you can start doing these use cases, and maybe it's just a few, but you're stacking up these use cases while you're working on the bigger problems like churn, audience growth, revenue acceleration, sales pipeline, velocity, Whatever those bigger problems are, that's the direct problem about people. They're trying to solve these bigger business problems with a smarter approach. And so I just see the roadmap as like, in a simple way, this is the future state. We are going to look like this three years from now. Tech's obviously going to keep evolving, so none of us can predict exactly what it's going to look like.

37:22But this is what we're going to do. This is our mission around this. We are committed to becoming this. Here's our responsibility of principles. And now we're going to start doing three pilot projects per business function per quarter. We're going to see them through. We're going to see if they work. And simultaneously, we're going to start doing one big problem statement per quarter per function. And so that's what the roadmap plays out is what are the resources, what's the team availability, what tech are we going to need, and then who's going to do this? And to your point, who's going to do this?

37:46I have no idea. You and I have talked about this before. The problem is if you buy into all this, and hopefully it makes sense and elements of this are at least, yeah, we got to do that. What I don't know how to tell you to do is who's going to do it. Because the reality is there's so few people in the world right now that actually understand this stuff at a deep level and also are business people. So there's a lot of AI people. There's tens of thousands of AI researchers and machine learning engineers. They're likely not the people that are doing this. They're technical people. They're building things and they're solving problems they're given.

38:23They are not the strategist, the visionary. It's a hard problem. If you can find a consultant who knows this stuff, don't let them go. Like they're, or if you have someone internally who can do this stuff, build around them. Because these are going to be really hard people to find for the next couple of years. There just aren't very many of them. Yeah, I agree a hundred percent. I think there was a totally unnecessary craze about the whole idea of prompt engineers. And this is not going to be a thing. But people who understand AI and business and strategy are very hard to find, especially at the higher levels.

38:57like people who can help navigate a corporate into that direction. There's not enough of these on the planet to solve the amount of issues we're going to have. Great strategists are hard to find. Like in the 16 years I ran my agency, it was the hardest thing I had to do was hire for people that I thought had a high ceiling as a strategist and then have the patience to get them there. Because you don't come out of college as a strategist. I'm sorry. It takes a mix of experience, of failure. You got to go through some stuff to become a great strategist, and you got to see some things. So we need great strategists who also happen to be extremely knowledgeable about AI and what it's capable of.

39:38It's a very hard hire. I agree. I want to start opening this to questions. There was already a few questions. I want to relate to something that relates to something that Paul Chaney wrote in the chat. And he said, I'm just a solopreneur. This is way bigger than my problems. I think the smaller the company, the bigger the opportunity and the lower the risk is right now. So this is going to really level the playing field, good or bad, right? If you're small, you'll be able to move a lot faster and do incredible things because you have lower risk and lower overhead and a lot faster to change. And bigger companies, especially regulated industries will not be able to react as quickly.

40:22And just to give a small example, and I don't know, I'm sure Paul, you've seen it, but just to the crowd, there's a guy, he's a computer developer. I think he's from Pakistan, but if I'm butchering it, then please excuse me. And his job was the thing, his niche was to develop applications for Shopify. And he would send these outbound emails to companies and say, hey, I'm a Shopify developer. I've done a hundred of these things in the past. Here's a few. And he would get an open rate of 0.01%. And every now and then somebody would actually hire him. And he went and scraped 100 ,000 Shopify stores.

40:56And he gave the data of what they sell, what they promote, how they write their stuff, who are their clients, all the data he had, he gave it to ChatGPT and said, what app would you develop for this business on the individual level, on a one business level? And then he had ChatGPT draft an email to that company and send an email. So he developed an application for himself that does these three things. Go research this company, define the app that will help them the most, send them an email. He sent 3 ,000 of those in the beginning and he got a stupid amount of replies of companies saying, this is awesome, this is the best idea I ever want to hire you.

41:35The crazy thing is he didn't know what he was offering because Chachi PT made it all up. The reason I'm telling you all that story is because this is what's possible today. And an individual person can do these kinds of things. But if you go to the next level over, which maybe is not 100 % possible now, but will be possible very soon, ChatGPT can write those applications. So now a company of one can be a machine of extensions for HubSpot or Salesforce or again, Shopify plugins or whatever the case may be, one person, because you'd just be managing this AI machine that will do all these things.

42:19So if you're not bound by the thinking of what was possible so far and just, oh, I just want to replace writing with AI writing, the world changes dramatically. So I really want to go to questions. And Kai had a question. Kai is saying, in some ways, is ChatGPT a distraction from longer-term enterprise solutions that could drive efficiency sooner in more secure environments? Yeah, that's a good question. I do often lead off my talks with ChatGPT is a blessing and a curse. It's just the tip of the iceberg. So it was a blessing because everyone woke up and realized they needed to care about AI. It was a curse because they thought they solved ChatGPT, they solved AI.

43:02So anybody who's followed AI at all knows, listen, AI was envisioned in the 1940s and 50s. This isn't new technology, this idea that machines could have human-like capabilities. It just took way longer than anybody thought for it to become a reality. And so there's hundreds or thousands of use cases for AI well beyond AI writing tools and image generation. And that's, I think, where we need to get to is this better understanding of what it actually is, what it enables to happen. And then from there, you can get organizations that are way better at figuring out and optimizing for specific use cases that are going to have the greatest impact on revenue growth or cost reduction are the two main grand goals people would have for using AI.

43:50Yeah, I agree 100%. I think it could be a very big distraction to my short answer to this, because it's the hype and it's the thing. And there might be bigger problems within your business that you got to solve. So don't make this be the kind of like the butterfly effect or whatever. No, look at this. And the shiny object kind of thing. It's something you got to deal with, but it doesn't have to be this. It doesn't have to be this, but it should be part of the things you're looking at. It's just, I would look at more of generative AI, text generation, and all the other things it's capable of, summarization, ideation, transcription.

44:27every company does that across every business function like that's why i think it's so obvious that what chat gpt made apparent is that this is going to be so widespread its use and its benefits that we're still trying to we're still figuring it out honestly like developing these understandings of its as a strategy assistant like i i've talked about that i think on the podcast like i i just don't think people have a clue how good it is at strategy like it is i've used it for four weeks at that. And I'm still, my mind is blown when I do it. So I think once people start realizing the true capabilities of these things, it's just going to be such a massive shift in the way work is done.

45:08And again, not even though we're going to lose a bunch of jobs, I'm just saying like fundamentally the way you do your work. I'll give an example that I think is super tangible. I have a buddy who's a dentist and he had reached out to me. He was talking about like the impact of inflation on costs and the inability to raise pricing because you're locked by insurance. And I was like, yeah, that's a problem. I don't really know dental businesses too well. So I went into GPT-4 and I just laid it out. And I was like, yeah, I'm a dentist. I'm locked in my pricing. Inflation has really cost me. What are 10 ways I can drive efficiency in the business to maintain profit margins?

45:40And it came back with 10 insanely good ideas. I sent them to my buddy and he's like, dude, three of these are legit ideas. I'm going to go check into this. I was like, there you go. I'll charge you for the minute of time. until we put it in a GPT for you. But here's the thing. He is an expert in dental practices. So I look at that thing and I'm like, I have no idea if these are even relevant. I don't know insurance. I don't know any of this stuff. He's the expert in being a dentist though. So as a strategy assistant, he can take something and do something with that where I can't. And that's where I think like designing, writing, however you're going to use these tools, if you're the expert in your domain and you know your stuff, you're going to get way more value out of them than the average person.

46:21And you're going to be able to just do more work. So I was a consultant for dental practices. I could take on 20 clients now. You know what I mean? Yeah. One thing, what you just said, the subtext of this, and we started with that, and then we'll go to one last question and then I'll let you go. I know it's been a long week for you, but the charging by the hour concept for, I don't care if you're a lawyer, a consultant, whatever you're charging, you're done. You're not done in your business, but you got to find a different business model because people will know that you can get this level of responses, results, documents, whatever the case may be in seconds or minutes instead of days and weeks.

47:01So you won't be able to charge 400 hours of a paralegal for something when people know this actually generates it in 35 seconds. So I'm not paying you for 400 hours. So one of the big, huge issues that the different kind of consulting services out there is the whole billable hours business model is going to have a very serious issue that it's going to be facing. The next question I want to ask is a question from Blake. Any thoughts on relying on cloud services versus investing in-house computing capacity for AI workload? I have no idea how to do the in-house one. The problem right now, so if anybody's not familiar how this stuff works, the general of AI stuff requires a massive amount of computing power.

47:44And the computing power largely comes from AWS, Microsoft Azure, or Google Cloud. And the infrastructure that powers those largely comes from NVIDIA. So the GPUs, chips that power it, and then Google has their TPUs. But there is a scarcity for those things. I know people within Google who can't get compute resources. So the The challenge is to do this stuff at scale, you need access to cloud computing and access to the GPUs. So if people within Microsoft and Google can't get access to their own cloud compute, it's really hard to think that other enterprises are going to be able to build up their own infrastructures to do this.

48:26If you can, great, but I don't know how long it takes to even get GPUs from NVIDIA right now. They may have a backlog of orders. So it seems as though it's a cloud solution you're going to build in these clouds. And what I've known recently is Google Cloud is giving significant credits to startups to compute degenerative AI in their cloud. So I think you're going to have a battle between Azure AWS and Google Cloud to try and win customers over by giving, and I'm talking hundreds of thousands of dollars in credits. if you're a big organization, probably millions, because they're trying to win you and lock you into their cloud for the next 10 years.

49:07So it's going to be a really fascinating time, but I think for most organizations, you're relying on a cloud provider, maybe a few of them. I don't know. I agree strongly. I think the only fear with the whole cloud thing is where's my data? Do I own it? Who has access to it? And I think all three organizations, which I agree with you 100%, they own the solution. Nobody else can develop the same scale of stuff. Already said that you can have a confined environment that is absolutely yours, that you can drop your data into it and that you can train your data on it and you can run it within that confined environment that doesn't leak anywhere else.

49:41Most of these organizations already own your data anyway. You're probably running all your data on one of those three giants. And so your data is there. Just training your models on that, running your AI stuff on top of it is just another layer of services they already provide today. And like Paul said, it may be cheaper for you to actually start using their AI services than just using what you're using today. Certainly faster to get up and running and get to value. But yeah, I think again, it goes to each organization. We're in a highly regulated industry with lots of governance over how you use data and maybe have personal information data.

50:17It's a different ball game, obviously, but you're used to that. Everything has barriers to it if you're in that kind of industry. Paul, I can't thank you enough. I think you and I can probably do this for the next two or three hours, but this was really fascinating, really well thought of. And again, you've been doing this for so long. I really appreciate you taking the time. I really appreciate you sharing all of this now, not just here on the podcast, but on your website, on the courses that you're doing for free, on the training you're doing regularly. If people want more of you, what's the best way to follow you, connect with you, work with you, and so on.

50:57Everything is housed at marketingaiinstitute.com, the book, the courses, the events, that's the blog. The podcast, as you mentioned, is that the listenership's been exploding, which is awesome. We love hearing from people listening to the podcast. And then me personally is LinkedIn. That's where I'm most active from a social perspective. And I try and share a thought once or twice a day, just put up kind of some observations about what's going on. So that's where I'm most active for sure. And I appreciate you having me. It's a great conversation. I love having conversations with people who are also really deep into this and think deeply about it.

51:29I think it makes for wonderful conversations and helps hopefully people are trying to figure this stuff out to start develop their own points of view, because my point of view and your point of view isn't what you listening should take, but synthesize it all and figure this out for yourself. And that's how we'll get ahead as an industry. Thank you. Yeah, I agree 100%. I think the number one thing that can help businesses survive this and society go through this in the least painful way is awareness and conversation. So I really want to thank you. Have an amazing weekend. Thank you, everyone.