Product of the week
(For a summary, click here.)
John Eischeid: Let’s start off with you giving a brief intro about yourself, how long you’ve been doing this – just kind of for the record – and then we’ll get into the book in a bit.
David Berkowitz: Sounds good. Yeah. I’m David Burkowitz, founder of AI Marketers Guild and a longtime marketer and a community organizer as well. Just done a lot over the years to bring marketers and others together. I really cut my teeth for a long time on the agency side and a mix of corporate strategy and marketing and innovation roles and then wound up going over to the tech side, working in the marketing tech and ad space and others – in either an in-house CMO role or fractional CMO role – and also a newly published author.
JE: Yes, and that was my next question. What made you decide to write a book about AI and marketing when it’s obviously something that’s changing so quickly? Are there certain timeless things about AI and marketing that are still going to be relevant 5 to 10 years from now? Because I did notice in the book that you actually kind of allude to and talk about online resources that are going to keep changing. So, just kind of flesh out the difference between some of the ideas in the book and what’s online.
DB: Yeah. This was the first question I was addressing with the publisher. It’s like, how do we – even with a pretty rapid publishing schedule, by the time I submit the manuscript, there’s the risk of things being outdated, so part of the strategy in writing the book was actually to mention specific AI models or tools as infrequently as possible. You can’t write a book about AI and not mention ChatGPT, but how do you just avoid getting in the weeds too much? And there are also things I’ve seen where, coming from this agency background, nearly 20 years ago I was one of the authors of this social marketing playbook when I was at the agency 360i, and I still keep it here. I’ve got the binder right behind me and the strategy around it’s really good even if Facebook might be the only platform in there that still exists under that name anymore. So what are those kinds of things – overcoming objections to deal with AI? They’re probably going to be pretty consistent now versus how they are a year from now, so if we get a step back from which exact tool to use, okay. There’s a lot that should be consistent, because it levels up to business strategy, and then yes, there are online resources. And I also host a community, too, if you want to know exactly if anyone using X to do Y, then you just got to talk to people, you got to go and see what’s useful in the moment.
JE: Yes, I did notice that you kind of shied away from mentioning specific versions of specific models. You just kind of talk about the capabilities in general. So yeah, I see that distinction there. You also mentioned in the book that you did use AI in writing it. You used it for research, but you wanted all of the writing to be your own. So, can you just tell me more about how you came to strike that balance between the two? What was your thought process?
DB: Well, there was the thought process and then there’s what I learned by writing it. The thought process is, “I want anyone to be able to go and copy a passage into some AI checker and the checker to show AI did not write this.”
JE: There are no em dashes!
DB: Yeah, exactly. You don’t see any rocket ship emoji all over the book. And it’s also, Why is the publisher putting my name on the book?” And is it really worth anything if AI even could write this kind of book? The part that I think surprised me more, and I’ve done a lot of writing over the years, the part that surprised me the most was I used AI to help me so much in writing the initial outline and writing the book proposal based on my own materials from presentations and articles that I’ve written. So, I said, “So, don’t just give me an outline for any book about AI. Here are some things I need for the proposal. Here’s what I’ve written and talked about. What kind of book would I write about AI and marketing?” And so, we got to a point – “we” meaning me in the AI tools – that I was pretty happy with, like, “I could write this book.” Once I started to write it, it was a whole different feeling, right? Because one of the challenges I had really early on was that it still didn’t feel as much like my work. It almost felt like I was writing to the test or what the teacher wanted to hear. If you’ve ever had an experience where you’re at a meeting or giving a talk and you’re using slides that you didn’t fully create. It could be your colleagues or things like that, stuff you’re familiar with, but then you just get to the slide and it’s like, “That’s not me. This isn’t my voice.” So it took a little bit of me working more myself in there, more of my anecdotes, more of my experience, more of my phrasing. I hope that people find a few things in there funny. And to the point where even if it could be done better technically if I had AI help me write more of it or if I stuck to the script a little bit more – like, “How do I just make sure my voice is all over this?” And then once I got there the book started getting really fun to write and it got way easier. And then I consult with AI more on things like, “Should this chapter come before after this one?” and just use it as a sounding board rather than a director.
JE: Kind of like, an editor on-call, or somebody, or a co-worker who you can just lean over to and ask a question of. So having said that, you did use AI for fact-checking, correct?
DB: Yeah. Yeah. Yeah. Which you then have to fact check.
JE: That’s what I was about to ask. How did you know that it wasn’t hallucinating while you were using it to fact check? Were you going through and using something like Perplexity with the links and checking the sources and that kind of thing?
DB: It’s funny. There were some things with the timeline that I had to go through a few times, because just the milestones in AI that got us to where we are right now. And there were definitely some debates out there online, “Was this published in 62 or 63?” like “When was this coined?” Something might have been announced one year but then released the next. So, it forced me to go and make sure I had a lot of clarity in what I was saying and how that was being presented. And sometimes it still relied on a human judgment call. But it was using that editorial judgment. There were some times also where it’s even just one of the most basic ones and memorable for me: I want to avoid jargon as much as possible in the book. So maybe the most common acronym when it comes to AI is LLM. Yeah. So when using this acronym, it’s like people don’t speak like that. Who the heck? Large language model? Learning? What is that? It doesn’t matter, right? So, I asked a few different AI apps, “Can I call this an AI engine? Is that accurate? And am I representing AI the right way?” And ChatGPT and one of the, Bard was Gemini by then, from Google. It’s like, this is actually a really good idea and this a relatable way. So just making sure that I was conveying things the right way was also really helpful because I have seen some of these things yeah
JE: You’re using the large language models to help you find the right terminology and to make sure that what you’re using is correct.
DB: Yeah. And if I was talking about a capability of something or I just wanted to make sure that even if I knew what X could do for Y, am I just phrasing this the right way? Is there anything that, like someone else could see, is inaccurate? And it’s funny because there’s another book, AI Marketing for Dummies, where I was the technical editor and I was using AI a ton to fact check their work and Perplexity was great for that and in going through the sources. But I also then became my own technical editor with AI’s help to hopefully save the editorial team some time or even be able to catch some things that someone who’s not a subject matter expert wouldn’t even know to look for.
JE: Could you give us an example of that?
DB: I think it gets into just some of the kinds of just capabilities and phrasing and terminology that I’m alluding to here and even when I get into the future talking about AI agents, something like, “Is this use case agentic?” Just making sure that for a book that isn’t meant to be technical, it still passes muster.
JE: You also include an AI readiness assessment as a questionnaire. I thought that was a really interesting idea. Where did that come from?
DB: For me it was a useful starting point. I try to have some content that’s a little more interactive, even if it’s in book form. I did work with Riddle.com for an interactive version that’s on my own version of the website but getting a sense of where you are with AI familiarity and what it can do, but also in a non-judgmental way. I would bristle if someone calls me an expert in AI marketing. Yes, I did write one of the books on it, but I’m just trying to learn a ton and try to go and keep pace with everything, too. So even me being more of like a translator for this and a curator and just someone introducing these ideas. I’m not getting 10 out of 10 on all these assessments. It was a good way to introduce some of what the book would wind up covering, without one of those chapters on here’s everything you’re going to learn in the book. I always skip over those chapters. I’m like I just want to read that.
JE: Right. You did include summaries at the end of each chapter, and I was kind of curious if that was for people who were kind of already very well acquainted with a lot of the ideas in marketing or some people who’ve been using AI and marketing already. They could kind of take a look at the summary and say, “Okay, is there anything in here that I might have missed?” and then read the whole chapter. Was your intention with the summaries to kind of move that reading process along or were you just trying to make it easier for people with maybe a little bit more of a background in the field.
DB: I think there are two use cases for it and the summary part of it is something that this is through Idea Press, which publishes these Non-Obvious Guides, so this part of their formatting as well. If I strongly objected to summaries, they might not, but with the approach to the summaries in this book, one is that if someone’s fairly familiar with it, they can just go and see okay, do they know those main points, or is there anything they should kind of go back and review? And the second is I also am always taking notes in books and I keep physical book report journals. I keep notes in the back of them and I actually write out the highlights. I can’t file a book away until I’ve written my mini book report in these journals. so having something that makes that part easier. It’s like yeah because often looking through I’m reading something on the Kindle, I’m reviewing all the highlights and I’m reviewing the popular highlights too and is there anything I missed that I should have caught? So I do find references like that useful.
JE: One of the things you do cover in the book is – you say that most people aren’t aware of what AI can do, and you even cite a cartoon from the Marketoonist about, “I think we may have a solution to all our problems,” and he’s got a hammer with “AI” on it, and a big bag of screws. I’ve used one of his cartoons in my LinkedIn feed before – with permission of course – but how do you think that we can bridge the gap between what people think AI can do and what it actually can do? And as part of that question is, “How do you address the fact that that gap is narrowing in some respects?”
DB: Some of it is just constant experimentation, if someone just reads this book and then doesn’t try anything, it’s worthless. And one of the other things that I do is through AI Markers Guild, I host a weekly call. It’s almost always a guest speaker and it’s just like trying to learn about how others are using this, where this fits in. And sometimes it’s really high level, like having Rashad Takawala, who was Chief Strategy Officer at Google, talk about where his work is going. Sometimes it’s very, very tactical and like, “Here’s something you can do with it today.” And the fact is that’s always changing, so one point – I bring it up in the book, I bring it up in literally every talk I give and I probably always will until I get bored of the slide – is that the most important word in AI is “yet.” And for the good, bad, and ugly with AI, if you’re saying, “AI doesn’t do this. AI can’t do this for marketing. AI won’t do this.” You have to add “yet” to it because it’s evolving so quickly. And that means the best cases and the worst cases like, we just haven’t seen them yet.
JE: Along the same lines, you do make a note in your book about a CAIO, a Chief Artificial Intelligence Officer, and how that might be going the way of the Chief Internet Officer or the Chief Web Officer of the 90s. Do you see any other parallels between AI now, the internet of the 90s, and for that matter, social media in the early 2000s?
DB: I see a lot of messiness right now. I think that there are some similarities in terms of just adding “AI” as a buzzword where it’s not needed. One of the things – I’m sometimes moderating panels being the interviewer – and I’m always asking people, “Okay so what’s different now than five years ago or three years ago?” because we did have a lot of applications of machine learning in particular. We didn’t have generative AI at scale three years ago, but we had other versions of AI being used. So, it’s especially for marketers, “What’s different for personalization?” “What’s different for predictive analytics?” And sometimes the answer is just much more of an evolution, not a revolution, right? So, having that reality check. AI doesn’t make everything better. It’s just like not everything needs to have AI vomited all over it. We still have jobs to do, regardless of technology. I think some of the differences are that from the business side of things, and if I talk look at CMOs and others of the C-suite and on down, the adoption rate has been so much faster. The same kind of brand manager, who in 2015 was still debating whether they should have a Facebook account or and in 2020 it’s like this Tik Tok thing it’s like just something for teenagers and isn’t really going to matter, so many of those folks were so quick to at least start dabbling into AI. And also I think that another really big difference is that we’re talking about the risks at a much earlier stage than we did with social media. If you look at books like Jonathan Higgs’s The Anxious Generation, and some of these teen cell phone bans, and increasing the age of consent – there’s a big movement globally. It’s happening in Australia. It’s happening all over the place of looking at some of the harmful effects of social media.
JE: Yeah, when I was in grad school, I wrote my master’s project on some of the drawbacks of social media and I was literally laughed at, not only by some of my fellow students, but even by some of my instructors.
DB: Not surprised.
JE: So, I know exactly what you’re talking about.
DB: When was that?
JE: 2010 was when I graduated.
DB: So, yeah. No one wanted to talk about that then.
JE: Yeah. No, it wasn’t even on anybody’s radar. Nobody was even talking about effects social media per se. I was basically doing a bunch of research on internet addiction, which people had kind of done some research into, and there were people who just couldn’t stop the “point and click.” And I kind of went into some of the situations where social media – people were unable to get off of it. Flame wars were a big thing back then. So I went into a lot of that and I’m just kind of wondering: Our attention spans have gone down and there has been some research recently to indicate that increased use of AI LLM dampens people’s critical thinking skills and apparently doesn’t take that long for that to happen. So yeah, it’s good that we’re proceeding with caution – I think – at this juncture and this early. I definitely agree with you on that point.
DB: Yeah, I’m still hopeful, like you. My favorite writer on the subject – have you read Sherry Turkle’s work? She’s written books, Alone Together and a couple others. She’s done incredible work on social media front of just what this is doing to people and especially how kids, teens are using this. And one of my favorite sections of one of her books was when she talked about how they were studying – I think they were like teens, pre-teens, like middle schoolers I believe it was – just going on a field trip in the woods for a couple days like ditching their phones and all their social skills were coming back really quickly. All the things that were dampening their critical thinking skills – we just have to use them a little bit, it doesn’t mean they’re dead. We’re just putting them on hiatus. So we don’t get too sucked down into that whole WALL-E, version of the future, which is one of those amazing movies that now I can’t stop thinking about.
JE: It’s one of those situations in which science fiction is becoming science fact. That’s not just true with AI, but with a number of other developments. But those are a different conversation. Moving right along, you do mention in your book about using AI to do research for brand presence. I also do some work in trademarks, like registering trademarks for people, and AI has actually proven very good to find use of a specific mark, which is a very similar research process to using it for brand presence. But having said that, in that situation how do you also hedge against hallucinations there? Do you go back through and check the sources and that kind of thing?
DB: Yeah a lot of it’s to check the sources, and I mean there even things then when you’re doing things – like you’re going from that to visual representations of it and then doing a reverse Google image search. Yeah, there are sometimes where I’ve seen some logos generated by an I’m like, “This looks way too familiar and way too familiar.” So reverse image search winds up being very helpful in trying to see what are the closest things out there to this.
JE: So let’s see . . . .
DB: How much more time? I’m like, back-to-back today.
JE: Yeah. Sorry, we’re a few minutes over. Do you need to go?
DB: I should in a minute or two.
JE: Okay. That gives me time for one more question, which is one I did want to touch on. There are a lot of posts about how much new content like ads, blog posts, and that kind of thing that people are capable of producing with AI. And so, given the fact that there’s so much more volume coming out . . .
DB: Yeah.
JE: What direction is this heading? Are we going to get to the point where people are just so saturated with advertising, marketing, and touch points and that kind of thing that they simply become desensitized to the amount of content that’s being thrown their way? I’m just kind of curious as to whether or not we’re flooding the market of people’s attention spans and where you might think that we are.
DB: I think the market’s already flooded. To me, the concept that I can’t wait to write more about and that I can’t stop thinking about is the tragedy of the commons. And it’s in this sense that where – you hit right on it – where we’ve got some finite pool of attention and when everyone’s trying to go and encroach more and more on that and way more than they deserve, and everyone’s just trying to hide like we’re hitting this point of overload. On the B2B front, on the marketing side, it’s gotten so easy to just fully automate all kinds of drip campaigns, even working video in and just the number of things that you can do with someone working a few hours a week – if that – and just setting up just massive amounts of volume and personalization and all kinds of multimodal touch points. You start adding in AI powered robo-dialers on top of this. And the flood is only going to get worse and search engines and other gatekeepers try to keep some of this at bay. But I guess it depends on the person in terms of whether we’ve reached the breaking point. But something’s got to give. It’s not going to be self-policed, because it’s too cheap, it’s too easy for everyone to do it. And I think the only saving grace is that at some point we’re just going to have enough of these “Mad as hell, I can’t take it anymore!” moments. And so that when someone – actually yesterday, I spent most of my day in in-person meetings around the city. And just having these personal connections and, “w”What can we really do together let’s just get off of this email slog and just all the slop coming our way. How do we relate as people? How do you show that there are actual humans involved in this process, right?”
JE: Yeah, the pendulum has already started to swing back a little bit.
DB: I think we’ll just put a premium on – essentially it’s like that artisnal version of it. Now I don’t think of myself creating marketing copy as impressive as someone who can build a table and you have some piece of furniture that can last 50 years. But in all of our crafts, even just little exchanges with a barista or something – thank god it’s not an app, you know?
