How Can ChatGPT & AI Help You Make Your Business Better? - Guest: Marketing Strategist, David Boyle

Today on Video Mojo, I talk with David Boyle, a true marketing expert who is also an authority regarding ChatGPT and AI. In order to make your business communications clearer, sharper and easier to produce, we cover some of the key applications that will help you leverage AI and how to approach AI in a way that will work as well as describing the entire landscape of, of AI and business. I really think that this is an episode not to be missed. Well, welcome David Boyle to Video Mojo. It's great to have you here across the pond. Are you where in the UK are you?

I'm in London, so it's dark and overcast and probably rainy outside.

I'm in Northern California, near San Francisco. The fog has burned off and we're very fortunate not to be too hot. Well, you're an expert with regard to both marketing and branding and positioning, working strategically as a data driven marketer for some pretty big brands. And you've also written a series of books about AI called the Prompt series to establish your authority. How would you position yourself as an AI expert? Where do you come from and how do you approach it?

Yeah, well, I guess I'd start by talking about brand side of it because that's the expertise that my co authors and I brought to this, which is 20 years of grinding out growth at brands using data, whether it's the music industry with pop stars like Snoop Dogg or Katy Perry or like tiny artists who actually some of which never made it anywhere through to luxury brands or books at HarperCollins or TV shows at the BBC. But we are brand people, we're really research people with data people with strategy people at heart. Like that's our expertise. And then as it relates to AI, I mean we've always been interested in technology and how technology can help brands on those missions. AI came along, as far as we're concerned, early December of last year and just blew our socks off in the power it had to help us do the things we'd been doing for 20 years. So we obsessed over how can we make it as useful as possible. And that's really where we got our AI expertise from obsessing over, doing things we were already good at, but using AI to make them quicker and easier and, you know, more fun.

And I'm totally with you on that, but I'm going to push back just for the sake of, of the argument because I, and you know, in one aspect that's certainly true. I talk about it as a runaway train and as a tsunami and with, with that comes a certain amount of danger. So those businesses that are afraid to use AI, how do you talk to them?

There's a great academic study I Now use from MIT. They gave professional writing tasks to 444 college graduate, college educated people and half them with ChatGPT, half with. And the results were clear. The group that used ChatGPT, I produce better work more quickly, with more self confidence. They had more time on the strategic parts of the work like editing and refining, less on the hard labor of like first drafting and ideation. It's just a list of positive benefits. These are my experiences, these are the experiences of other people, and these are also the experience of people in academic studies. So I think we're past the point of debating whether this is useful for people doing professional writing tasks. We're past that point now. Happy to have that conversation. But that's not a useful debate. The useful debate is how can I get those benefits? If you choose not to have those benefits, that's fine, they exist.

Good luck with that.

Sure.

And that underscores and I think that for myself personally, the biggest benefits do have to do with that ideation and first draft generation. And one of the things that I heard you say earlier in another presentation was that AI gets you 70 or 80% of the way. And I think that human element of adding that 20%, that you don't take AI's output as the gospel, talk about that process, because I think that that's key to people understanding how we develop a useful relationship, a productive relationship with these tools.

Yeah, a good way of talking about it is the mistakes people make. And I've worked with some incredibly smart people from some incredibly big brands around the world, you know, people in leadership positions. And I. One mistake I see people make, even the smartest people, is to do what you said, which is to take AI's output and then run with it. That's a golden rule for me. You should never do that. You should always edit, refine, fact check and enhance AI's output. An absolute golden rule. But it's a simple mistake people make. And so, so I think simple as that, you have to own the output, both in terms of authenticity and credibility, but also in terms of factualness as well. So for me that's an absolute golden rule. The best way to think of it is smart intern who's giving you very thoughtful work. But you would always check your intern's work and you would always enhance and refine it before passing it on. And so yeah, it is not something that you run with. You always own it.

And I, yeah, I agree that super smart intern is an analogy that I use. I also have heard, I always talk about it as a collaborator. I heard the term copilot used today. Yeah, that's awesome. So you're a data driven marketer, right? You guys are all about the data. And I think that that's wonderful from a scientific point of view, but I also think that marketing is changing. It's become less linear so that people who are thinking about funnels as 1 to 2 to 3. One of my other guests on Video Mojo said that it's like a pinball machine these days. People bounce around from place to place. So how do you track data driven marketing when people are operating like a ball in a pinball machine versus somebody that moves from A to B to C?

Yeah, first let me echo your sentiment there. I think the world is chaos and if you don't design for chaos, you're in big trouble. So people don't follow linear flows, but also decisions don't get made the way you think they get made in organizations. So organizations aren't good at good processes for decision making or for rolling out marketing anyway. So what we do to the audience is chaos and how they behave is chaos. So yeah, we're in a world of chaos. For me, data is most useful when it guides strategy. And so I'm much more interested in helping leadership to deeply understand their audience and the needs their audience have that are related to the category in question. That level of deep understanding then can cascade down through these chaotic decision making structures, integrate creative work or integrate campaign structures or whatever it might be. But planting those seeds in the head of leadership to really understand who their audience are and what their needs are and the kind of things they should do as a result. For me, that's the best possible use of data. So I'm less worried about campaign tracking and campaign optimization. That's great and really interesting. But if you get the strategy right, then everything else cascades down to be better, even through all the chaos.

In terms of making sure that brands and companies speak directly to their audience and address their needs directly, how does AI play into that?

Yeah, so for me it's all about some things you said earlier. So first, using AI as a thought partner and second, using AI as like idea generational brainstorming. And so AI, AI. If you think about what AI knows, it has read, you know, social media and philosophy papers and psychology papers. It has read audience research reports and theories and a hypothesis and it's read brand strategies. And so it's pretty good at understanding audiences and their needs and their wants and their hopes and their fears and their pains. And so you could ask it to be very thoughtful in your category. Let's say you run a concert venue, please, hey chat GPT. Please help me to think about the kinds of people who are interested in going to concerts of this type. Please help me understand their deep and underlying needs, not just their surface behaviors as it relates to to concert going. Questions like that will get you very, very thoughtful insights into who the different audiences are that you might target. And then chat or AI is more broadly a great thought, partners in bringing that to life. So questions like, hey, how might I innovate in order to better meet the needs of this one type of person? What product or experiences might I do? What types of marketing and communications might work for that, for that set of underlying needs? So for me it's all about what are the needs that are underlying that I can meet? And then help me think through the ways in which I can help. I can meet those needs.

Yeah, it's actually something that I think very few people have experienced. And I totally agree with you. There's this power there in ChatGPT I've had the most experience with understanding ideas like target audience and psycho demographics and all of that. So you've written this series of books called Prompt. Right. And you've targeted for real estate agents, for music industry for there's a half a dozen different markets. How does it vary from market to market? A two part question, how does it vary from market to market? And then how the heck do you keep up given how fast even just chatgpt is changing, not to mention the myriad of applications.

Yeah, I guess. 3 thoughts on that really. I mean the first is that the fundamentals of most industries are the same, especially if you look through the lens that I believe businesses should look through. Which is who help me understand the underlying needs of my audience, not the surface behaviors, the underlying needs. Help me work out what I do differently as a result of those needs and help me be creative off the back of that. That basic workflow works for everybody. That's why we can do five books on this topic. But the specifics of every industry are very different. What is a useful need? What type of actions might I do? What decisions can I make? They're very different. A screenwriter in Hollywood is making very different decisions than a real estate market in Florida. It's making different decisions than someone developing theme parks in the uk. You know, we've worked with all of those and the application of those standard theories and standard ways of working but to different people's different ways of working has been absolutely fascinating. And that's why we keep doing books, because we want to learn from them and the specifics of their industry and then help them with what we've learned from other industries. You'll see on the updates point, you know, we published the book late December of the first edition, and we totally assumed that we'd come back after the Christmas holidays and have to rewrite it and update it, and we didn't for about two months. And it turns out that if you operate at this deep level of really thinking about what can it do and how can I, and what are some good lessons for how I can make it behave the way I want to behave? Those lessons actually are universal. They're still true today. So there's different additional things you can now do. And some of the things we said it's not very good at, that it got good at. But the fundamental lessons for how do you think about these things as a copilot or as an assistant? How do you iterate and refine? How do you take ownership of its output? How do you improve a prompt until you get a result that you want? Some of those lessons, turns out they were true in late December when we wrote them down, and they're still true today. So we're lucky for that.

But it makes sense. It's kind of fundamental thinking. So I'm thinking about people who have not yet dived in to this soup that we're swimming in. And, you know, if you were a business person, like many business people who have not yet put ChatGPT or AI in general to work for your business, what's like the number one prompt that you would recommend to somebody to get started, to get a taste, to have a profound experience that'll get them moving.

Okay, I've got a different answer to this than I would have, than I would have given if you'd asked me this in December, actually. So here's how I would start. I would take the iPhone app for ChatGPT very specifically, or the Android app, which is now out as well. I would do the voice to text feature because it does really good voice to text on there. And I would go for a walk and I would speak into the app, some thoughts that you've got about your business. Hey, I think that we should be doing this kind of thing, and I'm worried about this and my staff. You know, here's a challenge here. And I've got this meeting on Thursday and I'm bothered just like share some thoughts with it. It could be coordinated and structured, but it doesn't have to be. And then end your prompt with hey, I run this business or I work in this business. Here are some of my thoughts. Please organize them for me, structure them for me, and recommend to me some next steps that I should do to kind of take forward each of the issues I talk to you about. You'll be absolutely blown away that with such a simple and engaging way of interacting, your slightly chaotic thoughts, which they inevitably will be, will be beautifully organized, beautifully summarized and beautifully elaborated on. It's kind of mind blowing. And I do that several times a day.

It is mind blowing. And there I did a similar kind of exercise with a Web app called heypi heypi.com because I like the voice activation and it speaks back to you, actually. Which brings up the subject of apps. And, you know, now that we've got ChatGPT plus Claude plus Bing plus perplexity, but on and on, you know, hundred, literally hundreds of AI driven apps. Do you stay with ChatGPT? Do you research the other apps? Do you have opinions about what's better or not?

Yeah, I do, yeah. I think for most people the best place to start is ChatGPT. It's a good, solid base and they're constantly innovating and it's. And it's kind of ripe for mass adoption. I think that some other models other than the OpenAI models are very useful. In particular, Claude from Anthropic is very useful because it allows you to put up to 100,000 tokens or about 70,000 words of context into it. So this unlocks new things where you can paste in like an entire report or an entire massive email thread or transcripts of dozens and dozens of qualitative interviews. And you can ask it right then to kind of summarize or take insights from them. So that's very, very useful. And to access that, I would go to a site called Magi M A G A I and what they do is bring together lots of models. So the OpenAI models are there and the Claude models are there as well. And there's some other features and functions in that side. They're just slightly better than ChatGPT and more useful. So yeah, once you've mastered the basics in a simple place and decided you've fallen in love with this as a partner to your work, then yeah, there are more sophisticated tools out there. They would be the two I use most. Yeah, GPT4 from OpenAI if I want like rigorous, thoughtful, structured work or Claude if I want more thoughtful, like better written work with longer context.

Yeah, that's very useful. And I like Claude too. I have to say. It's nice that Anthropic is a independent alternative that seems to have a forward looking culture. Is that your impression as well?

Yeah, definitely. I really like what they're doing as an organization as well. I think for most people they're the only two you need to worry about and you only really need to go to Claude if you need the extra context. In the meantime, Master ChatGPT is one place to go. Master that first. Our entire book was written purely on GPT 3.5. We've updated it now, but you know, that's how much great work you can do in just the old original chatgpt. So don't worry about the rest until you've mastered something, then move on and add loads of extra models to your repertoire.

Great. Well, two more areas I want to cover. One is the dangers. You know, I listened to a podcast recently with Scott Galloway from nyu. Prof. The Prof. G Pod interviewing Mo Gaudat. G A W D A T and Mo is a former Google executive. You probably are aware and they scared me that the misinformation and the bad actors taking advantage of AI could disrupt things as early as any time. But they're predicting actually early next year because of the US election. Are you worried about our future as well as the positive sides of the power?

Yeah, definitely. I mean, I think that there's a lot of different dangers to worry about. Absolutely. I said bad actors is definitely one. I think this is a superpower and it's a superpower that good people will use for good. But so I absolutely think that's a big, a big issue, a big challenge. We should worry about that. But it's more worrying about the bad people who will use it than AI itself then I think, you know, I worry about people making bad decisions off bad data. So I worry about people who don't fact check AI's work and don't take ownership of it and just blindly follow the recommendations it makes. That's wrong will lead to some bad decisions being made.

Don't do that, don't do that. People important warning, don't do that.

And yeah, I know senior executives who've done exactly that and they shouldn't, I've told them. And then AI, once it can control real world systems, AI itself might make decisions that are not in our best interests. And I Think that's something to worry about in the future as well. Very, very seriously. I think that one's worth worrying about in the same way we worry about global warming, which is we absolutely should worry about it. We absolutely should put resources into it and every day we should spend time thinking about and trying to prevent it. But also we go on a plane when we really need to be somewhere. And with AI we also should use AI to get our job done and be better, quicker and clearer in our work. But we should also worry about the long term benefits. It's not one or the other though. It's not an excuse not to use AI.

Yeah, no, I, I, it's interesting that you brought up global warming and climate change because it's also feels similar to me in the sense that it feels out of control and humanity as you know, our, our societies have to work together to get that done and I don't really see that happening. So I'm normally an optimist but there it seems out of control to me and yeah, but at the same time why not take advantage of it in order to do good work and to make our lives better and so forth. The way you're talking about, well, the other arena and I guess I'll segue off of the scary part. I saw a news report this morning, just this morning from India of Indian television using artificial anchors on newscasts. I mean literally this is not someday this is happening. And so I just wanted to ask you about the audio and visual world. I've been having a ball playing and creating graphics in mid journey and of course now we have Runway ML and other video things happening. So there's all those artificial audio and video and as well as generative AI, audio and video. And what's your view about where we're going there?

Yeah, I think that the text based AI is a great proxy and place to learn lessons that will then roll out to image AIs then video AIs and then music AIs probably in that order. And text based AIs are much more advanced. They're ready for prime time right now. I think the others aren't quite. It takes more work and more skill and more frustration to get the other ones to work just right now, maybe not next week. And so let's learn a couple of lessons from the text based AIs. I mean one, we've mentioned it roughly before but I think it amplifies your existing abilities and so for example in text like I'm not a great creative writer and it amplifies me to be like a half decent creative writer, but it ain't going to make me write me a novel. It's not going to do that. I haven't got the skill to use it in that way. Whereas somebody who's a great writer gets amplified and they become a truly exceptional writer, I think, with this. And so I think the same is true in images and the same is true in video and will be true in music as well, which is, you know, if you're a man off the street and you've got no particular skills, sure you'll do something that's fun, but it ain't going to be shown in a movie theater and people ain't going to buy it. But if you're a great artist with great skill, it amplifies that and gives you a real advantage. So worry about people who are less good than you using AI to get as good as you. Definitely. But don't worry about falling behind because once you use AI, you'll be amplified more than everybody else and you'll still be at the front of the pack.

And as somebody who's been involved in the music industry, what's your view regarding the copyright issues and other people, you know, people's work being incorporated into new AI generated work?

Yeah, I think there's a huge ethical dilemma here, which is, you know, every time I use a language model, I use them all day. Every time I use a language model, I'm very aware that although it's generating something that's unique, for me, it couldn't have done that without all of the copyrighted material it's ingested along the way. And so morally, some share of revenue or, you know, or value should be attributed back to the people who fed the model. His work, created it, and the constant new feed of original work, it's going to need to stay relevant and stay up to date as well. So we should be passing value back to original creators. Absolutely. I think, though I maybe have a controversial view on this as well, which is that these capabilities are so exciting for humankind. If you think about people in underprivileged parts of the world suddenly have access to this incredible intelligence, this empowering ability to speak at a level of the chief executive of a major corporation, you to think as clearly as them. It's such an empowering technology for people and it will make such a difference in people's lives. I do think you shouldn't be able to opt your work out of being part of the training set. That is definitely a controversial view. But certainly there's a moral dilemma here, which is that we're using people's original work and we're not crediting them or rewarding them for it as well.

That's very awkward and no easy solution there. Well, this has been fabulous, David and I know we've covered a lot in a short period of time, which I think is useful to the audience. Is there anything else I should have asked you that you really think people need to know or that you want to share?

No, I think maybe just to recap, this is transformative to how you think and communicate language models. And so even if you're skeptical about that, ask, like, hey, well, what if David's right? Maybe I should give it a go. And then I think just start somewhere simple and get really good at it. And ChatGPT is a good tool. You know, thinking out loud and talking your thoughts is a good place to start. It doesn't matter where you start. Pick anything and just get good at it, and you'll learn from there. But I would urge everybody to do that. I think if you don't, you're going to really, really miss out. Really, really miss out.

Do you prescribe to what a lot of people are saying in terms of AI is not going to take your job, but people who use AI will, 100%.

I mean, it's so transformative to my work. I tell you, I. I don't think I could hire somebody who. Who wasn't willing to use it and wasn't willing to get good at it. And if I was in a company with a team of mixed people, the people who did would be promoted much faster than the people who didn't because they'd be better, quicker, clearer, and they'd be happier in their work. What's not to love, you know? So I have a real problem with people choosing not to. I mean, I respect if you choose not to use a computer and you want to use a notepad. I respect if you want to write letters instead of sending emails. I respect if you want to walk instead of using a car. Like, I respect all those things. But you ain't going to compete with people who do. You're just not.

I totally agree. And I testify as well that I can't imagine doing my work these days without these AI tools. It's just so empowering and as you've said, fun. So if you haven't done it, people dive in. So your company is Audience Strategies and Prompt MBA is where people can find your guides and you're Most active on Twitter, LinkedIn or where.

Yeah, Twitter and LinkedIn. Absolutely. Hit me up, friend me, whatever works for you. On prompt, mba, the whole project is to help people. So the books are for sale? Absolutely. If you don't feel like you can afford the book, just. Just message me on the site right there. It goes straight to my phone 247 and I'll send you a code. You're welcome to the book for free. I just want people to use this technology. That's all we care about. If you. If it's not working for you, again, on prompt, mba, just message me there or Twitter and just say, look, it's not working for me. I want to help, okay? Because I want to learn from why it's not working. I want to help. So this is a project to help people to. To level up. So excellent.

I love that. And audiencestrategies.com is the consulting company where you can also take advantage of David's vast knowledge. This has been really valuable. I really appreciate you taking the time, David, and on behalf of Video Mojo, thank you and we'll stay connected and keep going.

Thank you so much. It was such fun.

If you enjoyed this episode, I want to recommend another Video Mojo episode that was with George Kaoh. We talked about how to use ChatGPT and AI in general to enhance your creativity. It was really a meaningful and deep conversation about AI and creativity. You might want to check it out. Thanks again for watching and we'll see you next time.