Whether we realize it or not, artificial intelligence (AI) has already impacted countless aspects of our everyday lives. And as the technology we rely on continues to adapt and change, those changes will inevitably shape the future of our businesses, our industry, and our society. As agency owners, it’s our job to stay ahead of those changes — even if the thought of an AI-driven agency is, understandably, a little intimidating.
Our guest on this episode of Build a Better Agency, Paul Roetzer, is not only an expert on all things AI, but he’s also someone who is dedicated to helping agency owners and other industry professionals. Paul teaches those agency owners and industry professionals how to pilot and scale AI technology so they can use it to drive efficiency, growth, and creativity for their business.
During our enlightening conversation, Paul demystifies the fascinating realities of artificial intelligence by walking us through what it really is, how AI agency tools can help make us better at our jobs, and why shaping the future of marketing starts with understanding how to use AI platforms, programs, and services responsibly.
A big thank you to our podcast’s presenting sponsor, White Label IQ. They’re an amazing resource for agencies who want to outsource their design, dev, or PPC work at wholesale prices. Check out their special offer (10 free hours!) for podcast listeners here.
What You Will Learn in This Episode:
- How artificial intelligence will — and already is — shaping the future of agency tools
- Why AI is NOT here to take your job
- Where you’re already seeing AI in your everyday life, and where you WILL see it in marketing in the marketing world
- What AI technology can do to help you make your agency smarter right now
- Why Paul says that you don’t need to go shopping for AI tools — and what to do instead
- How AI agency tools can be used to drive efficiency and growth for your business
- What Paul believes a “Next Gen” agency looks like — and how to become one
Ways to contact Paul Roetzer:
- Website: http://marketingaiinstitute.com
- LinkedIn: https://www.linkedin.com/in/paulroetzer/
- Twitter: http://twitter.com/paulroetzer
- Facebook: https://www.facebook.com/PaulRoetzer
- Paul’s Books:
- Marketing Artificial Intelligence Conference (MAICON)
It doesn’t matter what kind of an agency you run, traditional, digital, media buying, Web Dev, PR, whatever your focus, you still need to run a profitable business. The Build a Better Agency Podcast presented by White Label IQ will show you how to make more money and keep more of what you made. Let us help you build an agency that is sustainable, scalable, and if you want down the road, sellable. Bringing his 25 plus years of experience as both an agency owner and agency consultant, please welcome your host, Drew McLellan.
Hey, everybody, Drew McLellan here from Agency Management Institute back with another episode of Build a Better Agency. Welcome back if you’re a regular listener, and if this is your first episode, welcome to the future of marketing. That’s what we’re going to talk about today. Before I tell you a little bit about our guest and introduce him, let me remind all of you that we have two different AE Bootcamps that we teach every year. And we have the advanced AE Bootcamp for people who are four or five years into their career or more. And we get a lot of 20, 25, 30 year seasoned veterans at these workshops. So doesn’t have to be early in the career. I think we still have plenty to offer, even folks who’ve done it for a long time with the tools and tips that we have, and I think a new perspective on the job itself.
Anyway, that workshop is in Chicago. It is June 16th and 17th. And so you can register for that or the AE Bootcamp. So the one, I was just talking about the advanced AE Bootcamp, that really requires people to have four or five years of AE experience under their belt. And then the regular AE Bootcamp, that’s for people who are just getting into the business, been in the business for a couple years, project managers, account coordinators. Junior AE, somebody’s got less than four or five years of experience. That is August 1st and 2nd, also in Chicago. And you can register for both at the agencymanagementinstitute.com website. So if you get to the website, look for the navigation tab that says how we help. When you pull down there, you’re going to see workshops and you’ll see both the advanced and the regular AE bootcamp there.
I will tell you that we always have offered a money-back guarantee if you come to the workshop and it wasn’t useful, you didn’t learn anything, it didn’t serve your purpose. All you have to do is ask for your money back and we’re happy to give it to you. I’ve never had to do it, but we are happy to do it if you would like us to. And again, June 16th and 17th in Chicago for the advanced AE Bootcamp, August 1st and 2nd for the entry level AE Bootcamp.
All right. So let me tell you a little bit about our guest. He is a repeat guest. He’s been on the show two or three times, and always comes with such a generous spirit, willingness, a desire to teach. And so Paul Roetzer is one of my favorite human beings and certainly one of my favorite podcast guests. So Paul, until recently owned an agency in Cleveland and many of you know him from writing the book, The Marketing Agency Blueprint and The Marketing Performance Blueprint.
Paul’s got a brand new book out. A couple years ago, probably three or four years ago, he started the Marketing Artificial Intelligence Institute and also the conference MAICON, Marketing Artificial Intelligence Conference. It is one of my favorite conferences. It is brilliant. And I went thinking I was going to be intimidated because AI is so ever-changing and we’re all still learning about it, but actually I didn’t, I felt welcomed and encouraged and inspired to dabble in it even more.
So Paul and a guy named Mike Kaput have put out a new book that launches officially at the end of June, but you can pre-order it now on Amazon or your favorite bookstore called Marketing Artificial Intelligence, AI, Marketing, and the Future of Business. And I was lucky enough to read a preview copy and be able to leave a comment on the book jacket.
And I got to tell you, the book is spectacular. It’s like a guidebook, just walks you through all of the different ways that we as agency folks can and should be playing with AI. And so it’s a fascinating read, but it’s a fast read. And I think it’s like an action-packed book in terms of getting you to actually do something, which I love. So without any further ado, I just want to introduce you to Paul, welcome him back to the show and dig into the content of the book. All right, let’s do it. Paul, welcome back to the podcast. Thanks for joining us again.
Always fun to be here. I don’t know, it’s been a few months, at least since we did this, but I’m really looking forward to the conversation as always.
Yeah. I think it’s been about a year. I think it was right before last year’s summit.
That summit, your event feels to me like it was so recent, because I was actually talking to Jay Bearer just yesterday, and I was like, “We just saw each other at Drew’s event?” And he goes, “Yeah, that was a year ago.” I’m like, “Oh my God.”
That wasn’t really a year ago. It was August.
Okay. Oh, it’s all right. So yeah, it doesn’t feel too long.
No, no. In some ways it feels like it was yesterday in other days. As we’re gearing up for the 2022 edition, it feels like, “Oh my God, we just did this yesterday. Why do we do it again for?” Here we are. So you’ve got a new book coming out this week. So tell everybody a little bit about that.
Yeah. It is Marketing Artificial Intelligence, AI, Marketing, and the Future of Business. And I co-wrote it with our chief content officer Mike Kaput who’s been with me for 12 years or something. We were together at the agency and then he came over and moved full time at the Marketing AI Institute. So he and I have been at this since 2016, writing blog posts doing talks. We’ve published probably 900 articles about artificial intelligence since 2016.
And so really this book was something I started on 2018, tried to write this story and just couldn’t, we knew where AI had been and we thought we knew where it was going. And in 2016, 17 we realized we don’t know the middle part. We don’t know how to actually tell people to get started, where to go, which vendors to use. We didn’t know the actionable part yet. And so we really had to step back for a few years and just grind and research and write and talk to vendors. And that’s really what the book is, a culmination of all that research, all the writing, all the demos and trials and everything we’ve done in trying to make this very actionable guide for people that want to pilot and scale AI.
So I had the good fortune of reading the book already. And gosh, I learned so much from it, but it certainly is structured in a very specific way. So talk to us a little bit about how you structured it and why you structured it that way?
Yeah. Our research has told us that the average marketer and certainly the agency owner, agency professional, really is it a base level at best of understanding what AI is? So it was very important to me to not retell the AI story. It’s been told very well. The history of AI has been told very well in a number of books. But I wanted to tell it in a way that made it make sense to marketers and agency professionals, where they realize the significance of what’s happening.
So the book starts off with the science of making machines smart, and it tells this broad picture story of how the major players like Amazon, Microsoft, Google, how they have been working on AI for decades. And it lays the groundwork like these tools, these companies that integral to, especially Google to what we do as marketers and search, just organic search on its own and paid that these companies have believed AI was going to change everything for 20 years, and they’ve had these stops and starts.
I tell this story a little bit of Microsoft with language technology, they’ve been working on this stuff, and it finally around 2012 became possible to commercialize and build products around the things that had largely been theorized in academics for decades. And so I wanted people to realize AI isn’t a new thing, but we didn’t want to repeat the other stories. So we wove it through the stories of those three major companies and then showed what their AI units look like and how many thousands of people they have working on AI at these companies, how many pre-trained models they have. So the basic premise was it is already here. The major companies that are driving the economy, driving our industry, driving business, they have believed this for a really long time and now it’s coming to us-
Yeah. So then it goes into language, vision. Prediction is the category. So understand big picture what these things are. Then we teach a little bit about how to buy the tech, but then the whole middle part of the book, we break into 10 categories of marketing from advertising and analytics to social media and email marketing and advertising. And each chapter is structured very specifically to introduce like, “Okay, here’s AI for advertising. Here is a sample case study. Here are sample technologies. Here are sample use cases. Here’s how you can get started.”
So if you’re an advertiser or if that’s mainly what your agency does is advertising, you can read the first four chapters, get a really broad macro level understanding of AI and then you can just dive in and pick the middle chapters of like I do communications and advertises. I’m going to focus on those two. And then the end of the book is more, what does it actually mean to you? How do you apply this to your career? How do you get a head start? How do you differentiate your business based on it?
Yeah. As I was reading it, I was taking notes and a lot of my notes were, I felt more confident about experimentation after each chapter, was like, “Oh, okay, now I get what this does. I have potential partners to experiment with. And I now actually know the questions I should ask or the things I should see if we can do.” I think the book empowers people, as you say, everybody’s at this baseline, it empowers people to begin to experiment without having to guess where to go or what to say or do or what to expect. So I think it is a field guide of experimentation, I think.
Yeah. And that was-
That felt to me anyway.
Yeah. And that’s by design because you do, once you break down the barrier of, “Oh, okay, this isn’t sci-fi, it’s not going to automate away all of our jobs. It’s actually just a practical technology.” And the thing we always tell people it’s just smarter tech. If your agency already does email marketing and you already do advertising and you already do social media management, you just need to be buying technology that has AI built into it, that makes you better at your job every day, that tech that itself learns and evolves. So when a company that you’re using introduces 10 new features, your agency doesn’t sit around not using those features because no one on your team’s been trained to do it yet. What if the software just gets smarter and continually new makes you better at your job?
And so that’s really what we teach people. And once it’s like, “Oh, okay, fine. I’m not scared of smarter technology. Tech’s always getting better. Now, what do I do?” Then it’s like you got to find something the trial. And I always tell people like, “I can sit here all day and explain to you that it’s not that scary. It’s not that hard,” but until you go test a GPT-3 tool and realize like, “Oh, it’s just assisting me in my copyright. This isn’t going to take my job as a copywriter. It’s just recommending some creative to me. That’s cool. It said it better, I hadn’t thought about saying it that way.”
That’s really what the vast majority of AI is going to be, is assistive in your life. It’s just going to help you be better at your job. And then it’s just like test a few tools and you start to realize like, “Oh, this isn’t really that big of a deal. It’s actually is just smarter technology. This is cool.” But you need that on-ramp for people. And so many people avoid taking that on-ramp because the topic itself just seems so abstract and overwhelming.
It’s interesting as I was reading I was thinking, we almost do some of the tech a disservice by labeling it with AI because that makes it scarier. As opposed to someone saying, “Well, go to Grammarly and you plug in your text and it tells you where you need commas.” That seems super practical and easy and… Well, there’s not a person on the other end suggesting where there are commas and things like that. But for some reason, the umbrella term AI, I think is intimidating still to a lot of people.
It is. And that’s, we try real hard to break down those barriers. Again, you had a great example, Grammarly, auto.ai. People do Zoom and then they get the automatic transcription. You don’t stop and think like, “Oh, thank God for natural language processing and generation that it just was able to take speech and turn it into text.” But there’s all these underlying functions, well, I’ll often say your life is AI assisted and your marketing will be too. And what I mean by that is Netflix recommends shows and movies, Google Maps recommends destinations or routes to get you to your destination faster. Gmail finishes your sentences by predicting your next word. It’s everywhere in your consumer life. Every time you get personalized promotions and stuff, you’re seeing it all the time.
The marketing industry, marketing technology just isn’t that advanced yet. It’s not ubiquitous and every piece of software we buy, but it will be. Five, seven years from now, you’re not going to be buying marketing technology that isn’t infused with AI. But right now you have to seek it out and you got to find the tech that actually is more advanced. Because we also say all AI isn’t created equal, just because that vendor buying a content strategy tool and one of them says they use machine learning for X, Y, and Z, that doesn’t necessarily mean that it’s actually that much better than what you’re already using. So AI isn’t like, “Oh, okay, it’s got AI good.” There’s lots of different levels of AI and how advanced it might be.
So I think part of it too is the industry and the tools are evolving as we begin to experiment and involved. So I suppose it’s a little different than walking into a mature industry or a mature offering, there’s experimentation on both sides, right? There’s still experiments. You were saying to me before we hit the record button hundreds of these companies are cropping up all the time and there’s a lot of new technology and a lot of new players in the space.
So everybody’s experimenting all together, which probably also makes it a feel a little more… When you say use Google, everyone goes, “Oh, I got that.” Right? We know what it is. It’s established, it’s mature. Great. But some of these other tools, I think, for some folks it’s like, “Well, are they even going to be around in a year,” et cetera, et cetera, et cetera.
Yeah. What we always tell people is, “You don’t need to go looking for AI tools. You don’t need AI.” What you do, and there’s basically two ways we teach people to adopt and from your agency. And again, I owned an agency for 16 years. I sold it last year, but the way I would look at it as an agency is the fastest path is to make a list of all the things your team does. So let’s say that 80% of your revenue comes from advertising, social media and email marketing. I’ll just make up categories. I would break them into those three categories. And then I would itemize all the tactical things we do. So for advertising, we develop creative, we test creative, we manage ad spend, we develop performance or points, whatever it is, just get a list of use cases.
Then I would put how many hours a month do you spend doing each of those things and how much money are you spending on the tech to do those things? And what you’re doing is you’re building a prioritization of where could AI potentially help me the most? And you’re trying to find use case say, “Okay, we’re spending 120 hours a month developing performance reports for our 30 clients about ad spend. And it’s all human, all manual development. And it’s being done by junior associates, doing what they’re doing.” That’s a use case. So it’s AI for performance reports. That’s what you need AI for. You don’t need AI. You need AI to do a thing you already do more efficiently.
Yep. Yeah. It seems so simple when you say it that way, but I think you’re right. I think people are still floundering around trying to figure out how to use it. So I know one of the things that you did was you actually used AI to write the book. So can you talk a little bit about what use cases you identified and how you went about finding the tools and how that all worked out?
Yeah. It’s a cool story that demonstrates for everyone where we’re really at with AI. We started development of the book last March. We started working on the idea, building the outline, starting to work on the manuscript and the research. And so I had this grand vision. I was going to redefine how book writing was done. We were going to put AI into everything. And I have this sandbox of all the things I wanted to use AI for. So one practical example was I wanted to use AI to write the summaries at the end of each chapter. So feed it the 9,000 words from the chapter and have the AI write a summary of it. And I’m not saying pick 10 excerpts like verbatim. I wanted it to write a summary, that tech didn’t exist.
So when we tried to do it, this content summarization, it’s called abstractive content summarization to be specific, didn’t exist, that where we could buy it off the shelf and do it. You had to pre-train on 10,000 models, 10,000 examples to create a model to do this thing. So I was like, “Oh, okay, well, I guess we’re not doing that.” We thought about having a synthetic version of me do the audiobook. You can just do Google Wave. People aren’t familiar, Google has Google Wave technology where if you invest couple weeks, you can actually train a synthetic version of your own voice and then you just feed a text and it’ll read it, and it sounds exactly like you reading it. So we toyed with that, we weren’t allowed. The book-
What do you mean you weren’t allowed?
The audio book industry isn’t a fan of the idea of voice actors being replaced by synthetic voices. So we couldn’t do it. I don’t think we could license the audio version of the book that way. So we just couldn’t do it. Now, we may experiment with it. We’re still debating on taking excerpt from the book and training a Google Wave model and actually doing just a demonstrate the technology. So there was a bunch of areas where I wanted to do, we couldn’t, but I will tell you where we did do it.
So to start, we’ve published about 900 articles on the institute blog. We’ve probably done about 40 or so webinars. I’ve been on dozens of podcasts. We’ve done dozens of podcasts. We have all of this audio, video and actual written text, combined hundreds of thousands of words. So the starting point was anything that we had done in audio or video that wasn’t previously transcribed, transcribed that, turn the speech into text, that is an AI application. So whether use auto.ai or Descript is another tool we would use, you could use a Google or Amazon, AWS, but we took every spoken word and turned it into transcribed content. Then we were able to actually look at that content and start building out the brief. And so we took chapter by chapter, dumped all this transcribed content into it. So something that could have taken weeks, we did in a couple days. And all of a sudden to write a 60,000 word manuscript, we had 200,000 words already there and you almost… It’s like the sculptor where you start carving [crosstalk 00:19:25].
So that was a starting point. Another would be research briefs if we had a topic that we were writing about that we hadn’t previously written about, you could use a market news or a phrase, a tool like that and develop a brief. And then rather than us Google searching 15 different times and copying and pasting links, whatever, just put a topic into a research brief generator and it would develop it, recommend what questions to ask and answer, things like that. So same kind of tools we would do used to develop our content for the site, for the institute, we used as the starting point for the book. And it saved us probably dozens or hundreds of hours off the hold.
Yeah. So one of the interesting things I think is that, and I’m reminder to all of us is that this is an immature technology, it’s been around for a while, but there’s still a lot still being developed. You probably know more about this topic than just about anybody, even you were still discovering what you could and couldn’t do.
Yeah. All the time. So the way we explain this is once you understand what the technology is capable of at a high level, then you can look at problems very differently. So in the case of the book, we were almost doing more problem identification like, “Hey, how do we do this thing? We have to create this thing in four months. We have to write these 60,000 words and do all this.” It was like, “How do we do that differently than we previously would?” And it wasn’t existing use cases. It was like, “Well, let’s reimagine how to efficiently create a really high quality book.” And so I think that’s where we’re at in the industry, is we don’t have enough people who know what to look for, to understand what’s possible.
And that’s our hope for this book, is we can just create more marketers, agency professionals, business leaders who understand what the technology is capable of and who almost challenge vendors to create these smarter solutions. There are some very established marketing technology companies that I would consider significantly behind the curve of building smarter technology within their platforms. The opposite of that would be an Adobe. Adobe in 2019, I watched a keynote on product features they were running out, they had 75 to 100 AI powered features built right into the platform. Three years ago.
So there are some platform companies that have seen this coming and have invested billions of dollars to build these capabilities, to make products smarter. So all