BCR 134 - Gary Cooper, AI specialist copywriter, France

Speaker 1 - Lee
Gary Cooper's an AI lead regeneration and marketing specialist, and we like to think of him as a creative catalyst, a tactical thinker and an AI Weaver. Gary, tell us a little bit about your background?

Speaker 2 - Gary Cooper
Yeah, sure. I got about 30 years in the copywriting industry, and I started using AI when phrase came out and Quill bot about four or five years ago, and I got quite into the the technology behind it. So when they released chat, GTP and Claude Gemini, all of those, I was already prepared for what it did and the ways in which it could help me. So I've just adapted and followed on through the system. Really, I think that the building blocks of the initial AI products that came out were very useful because they gave you a foundation. I find now that people get very confused, because there is so much out there, whereas back in the day, there wasn't, and you could cope

Speaker 1  2:14  
with it. Yeah, I totally agree. I think, yeah, it's like that, that comment that you made on the wranglers group today about how sad it was that that woman, who he's, you know, got tons of experience and expertise, is going to stop publishing stuff, because what she thought was she was, you know, talking to lots of experts, but that's not what people really want now they there's so many people coming on board. They want beginners level information, not the expert stuff. The experts are already, you know, running with it. That's right.

Speaker 2  2:52  
Then I think that that's quite a rarefied atmosphere to teach to. I think that the thing about the teaching and the education is that the beginners are, if you like, where the money is for for people, and the people who are into AI, for the AI, aren't particularly interested in going, if you like, back to that level. So it's, it's kind of separating into two camps. I think it's quite sad in in many respects, because, for instance, with with her newsletter, it was really good because she gave you a different point of view and understood all of the the technical aspects of it, which gave you a completely different insight. You can't do that with beginners.

Speaker 1  3:41  
So hence agencies like ours that are able to sort of bring guidance and influence and avoid all the technical stuff that most leaders don't really want

Unknown Speaker  3:54  
to get their hands involved with.

Speaker 2  3:57  
Yeah, I agree. I think that the problem is that people who are new to aI have read all of the PR on it and think that it's capable of doing everything straight off. They don't realize that there are different models for different purposes, different prompts for different purposes. The stuff that you use for graphics is completely different from the stuff used for text. It's completely different from the stuff used for programming. And that causes, I think, the most confusion the idea of AI for me is not to replicate what I can do anyway I can write, and I have written for years, so it won't be replacing that. What I look for AI to do is to do the bits that I hope doing that all of the admin, all of the strategy planning, all of the spreadsheets and all of the stuff which takes me away from my writing. So it's creating time for me to do other stuff,

Speaker 1  4:58  
no matter how good. Claude is there's just that little bit of human element, the humanity of us, the soul of us, that goes into our own copy, and it can't be replicated in you know, no matter how good Claude is, for example,

Speaker 2  5:13  
I agree. No, absolutely. Gary,

Speaker 1  5:17  
you'd firm. You obviously don't have a French accent, you have an English accent. So how has your experience living in France now influenced your perspective on AI adoption in different cultural contexts?

Speaker 2  5:30  
France is an oddball. I know everybody thinks that, but it's particularly an oddball when it comes to this, because they're very cautious. They are not very quick to adopt this kind of stuff, and when they do, they go all in. So it's reached the point where I think that the interest has now been generated, and people are beginning to look at it over here. But the EU are very strict with their rules. They brought in the AI Act back in February, and like all the EU rules, it takes a time for people to assimilate it and then actually bring it into the workplace. So that's being looked at. I mean, I know that in the Nordic countries, they've already adopted AI for most things, for Finland, Norway, Sweden, they're all very up there with it. But France is lagging behind in it, and the French are naturally suspicious kind of people, and because most of the tech aspects of it are American based, they are a little bit reluctant to use it. I don't know if you remember years ago, they tried to get a French version of Google because they didn't trust Google and they couldn't make that work. So it's a strange country in lots of respects, from that point of view. And then, of course, there's the language. Language, the challenges of AI in multilingual environments is quite difficult. So international business, the lingua franca is usually English, which the French don't like. So there are lots of things which are quite difficult and challenging. I go back in my

Speaker 1  7:29  
in my brain box a little bit, and I try and remember there was something I heard somewhere, and it could be complete nonsense. No idea, but it was that somewhere, France has got a lot or French has got a lot less words in it than the English or British vocabulary. Yeah. And so that makes it difficult, if you, if you're like an engine of some sort, to understand what a particular word means in a phrase, if that you know word could have six or seven meanings.

Speaker 2  8:01  
That's right. For instance, if you go to the doctors, there's only one word for pain, whereas in Australia and English, English, you've got sharp pain, got dull pain, got throbbing. Pain doesn't exist in France, it just hurts in France, wow.

Speaker 1  8:20  
Yeah, stunning, yeah, that must make it so difficult for can you share a specific example of how you've successfully integrated AI into the copywriting processes that you've got, and what impact it's had on the results for you and and also, if you've done that for your clients? Yeah? Sure.

Speaker 2  8:42  
Off the top of my head, there's a company I work for in California who do podcasts and they research questions, which the AI can do for me in perplexity, I use to do that. So they'll give me a subject that they want to explore on the podcast. I will go into and ask it to provide me with 20 questions on that subject. I can then go into perplexity to research the background of those subjects, and it will give me links and references, I can then go into those links and references, and I set up a project in Claude specifically for these podcasts, where I can analyze articles from newspapers, or I can analyze transcripts from YouTube videos or other videos and get short summaries of them. So I can ask it for, say, 12, 2015, talking points from any particular article and a brief two paragraph summary and the link, and I can send that to the client so they've got everything in front of them. And. That used to take me perhaps a day, and I can do it in an hour and a half now, which is staggering the difference, because just physically looking up all of those references would take you a long time, yeah, and then reading through them and doing your own manual summary would take a long time. So that has saved me. I reckon that's all of that type of thing that saved me, perhaps three days a week in terms of work.

Speaker 1  10:29  
I do a lot of writing of books and things as well, and they because I trained in psychology, so I quite often write self help psychology books and whereas you know the research that you have to go through, going through the databases and everything you know manually and search term now, I just go to consensus the the app and type in, in English, what I want, and it goes away and finds All these references of journal articles that are absolutely on point, and it's just, and it's just like you spit it out in two minutes, whereas I'd be spending a week, you know, in the databases at university. So it's a godsend.

Speaker 2  11:15  
It is indeed, and for research on that level, it's a godsend as well. Oh,

Speaker 1  11:21  
yeah, because consensus spits out a summary of each paper as well. So and if it spits out papers 10 at a time, it will, at the top of the page, give you a condensed version or condensed summary of all 10 papers, of what those papers are trying to say and what they mean. And it's just like I could not do that in a week. It's

Unknown Speaker  11:41  
phenomenal. No, no. Exactly.

Speaker 1  11:45  
Another question, what do you see as the biggest misconception CEOs have about AI in business, and how would you address it? I

Speaker 2  11:55  
think there's a big worry from the point of view of CEOs as to whether this is going to replace the human elements and their workers completely, whereas I think that I see my job, if you like, as having to explain that the reality is that it's a tool to augment the jobs that we all do and enhance the human capabilities, not replace them, as we already said in terms of research and that kind of thing. It's a time saver. And, yeah, okay, if, if you are a researcher, you've got far more time to actually go in depth to a subject, because if you're on a deadline, the biggest problem is things get missed or ignored or forgotten. So this is a way of actually catching up with yourself and enabling you to get more work done in less time. The other thing, of course, is the complexity and the cost of implementing the AI solutions, because when you get onto very, very deep problems, company wide problems, strategies, all the rest of it, it can become very, very expensive, programming, that kind of thing, development, All of those things, deep computing power, yeah, it's not a plug and play solution. You also need training. You can't do any of this without learning how to do the work properly, knowing how to prompt properly, knowing how to prompt for the results that you want. And I can see that that would be a problem for CEOs, because the people will be off doing courses, and that's a time suck. What they need to look at is the long term for this, and realize that a week spent on a course may well save them a year during the course of that, or whatever. It's something that, if you like, it's an upfront investment in both cost and time for a company to gain in the long run. I remember

Speaker 1  14:14  
reading a Forbes article from about two or three weeks ago, and it was American focused, of course, because it's an American publication. But the The upshot was that if you want to earn phenomenally big money as a contractor, be a prompt engineer. And they were just like just piling in the money you buy your own island after a year, because it's the power of the prop. I mean, we I go back to the because I'm an old, you know person. I go back to the 70s. And I remember in the 70s, computer coders lived and died by a particular word, and that was called GIGO, G, I, G, O, garbage. In garbage. Yeah, yeah, yeah. And. If you wrote, you know, appalling code garbage in, you get an appalling results. And as I say, I write books on in my spare time. And I am amazed by how many people come up to me, you know, when I've got a book, and they say, Oh, it must be really easy to write a book on artificial intelligence, because you just go to the, you know, chat GPT and say, write me a book. And you and I know that that's not how that happens. And so to bring it back to the example of, you know, a guy going on, a guy or girl going on, a one week workshop of learning how to prompt, if they come away knowing how to prompt for the result they you know, they want to get. Um, that's, you're quite right. That could save the company millions of dollars, yes, and especially if they put in little twists, like you taught me once, put in little twists about like, what is it about this topic that would surprise people, or what is it about this topic that most people don't know, putting little twists like that into the prompt suddenly takes you into a new area, and you go, Wow, I did not know that my client's probably going to want to know that. Okay, looking ahead, what emerging AI, technology or application, are you most excited about for its potential to transform business operations.

Speaker 2  16:22  
There are quite a few. I think customer service is one which I'm personally very excited about. The chat bots that we all know and love are a bit clunky and hot and you're going to like my bank yesterday, I'll go in and ask them something, and it will perhaps give me four or five options, none of which are the ones I want and none of which it can answer. And so I have to hang on to talk to a human and the whole thing was a complete waste of time. And what I love about the AI, natural language processing advantages that are taking place is that they can now they've reached the point when they when they get all the integrations together, where you can actually talk intelligently to if you like a person at the other end, which is actually an AI, and it will, using its artificial intelligence, actually be able to assimilate what you're asking rather and answer you. It will take into account all of its database of past things. So, for instance, you might phone up to ask about a particular paint that you ordered that came that was the wrong color, and the AI will look up its database. Oh, yeah, you wanted black. What have you got? I ended up with red. Okay. Well, look, send it back. We can get the White House to do you a credit note, and we'll get the other stuff dispatched off to you straight away, that kind of thing, very simple, but it's not the dumb type of answers that we've been used to. And the other thing that that brings on, of course, is the predictive analytics in the supply chain, being able to manage things like that. So keeping the just in time stock, very just in time that that process of the customer service could be linked directly to the warehouse so they know that they've got this paint going out, and they will have to reorder it without any human intervention whatsoever. It would just add it to the next order of the paint supplier. Very clever. And also, of course, the innovation and product development. That's another area which is just fascinating. I was reading a article about pharma company in Switzerland that are doing cancer research, and they can now simulate the effects of a drug on somebody's DNA, so that they can match the exact dosage and strength of the chemo that they're giving somebody to the cells of the cancer that's grown. And the accuracy is astonishing. So you don't get all of the side effects, but it's the most efficient way of doing it. They're also experimenting with gold nano particles because they've realized that you can heat the gold but it doesn't react with anything within the body, and the heat destroys the cancer cells without destroying anything around it. And the combination of these two treatments is supposed to be the results of phenomenon in on the AI simulation few years

Speaker 1  19:53  
ago, that was the Holy Grail for for pharmaceutical companies to be able to make a. A drug that is tailored exactly to you, too. So you know, your drug would be different from my drug, even though it's got the same name on the box.

Speaker 2  20:09  
Yeah. But again, it saves so much time and money on clinical trials, because they can go straight to the patient and try this out and, well, it's just mind blowing. The whole thing is mind blowing.

Speaker 1  20:26  
But, you know, to bring it back to the customer service stuff you were talking about earlier on, I can see a position where, as a customer, I go into my bank's website, I have a conversation with a person, a bot. That bot not only knows all the company's, you know, database, but it also has got me as well, my bank account, details, my transaction history and everything, so it can engage with me rather than it's just a generic please press one. Please press two. It's just like, okay, Lee, what do you want to do today? You've got this amount there, this amount there, if you do this, then that will trigger that result. And so you can make an informed choice, and you've not spoken to a single person

Speaker 2  21:13  
Exactly, exactly. One of the things that I'm working on at the moment is with dead leads lots of companies have 1000s of leads in their database that they'll never get around to approaching, simply because they don't have the staff do it. They it's very expensive to get call centers to ring and follow up on leads that are quite old. So, for example, if I have a company around to look at double glazing, so the panels, HVAC system, air conditioning, or whatever, I might want to get three quotes, four quotes, I might want to think about it, because it's more expensive than I thought. So the company will put that down on their their CRM or their spreadsheet or whatever they use, and it just goes into a computer file somewhere, and that's it. The AI can then come in and just say to them, fire an SMS message. Hey, look, three months ago, you've got a quote from us or an HVAC system. Are you still interested? And customer says, Oh, actually, yeah, I am. I'd like, you know, to follow that up, and the AI can make an appointment with the salesperson. Or if the guy isn't interested anymore, they'll just turn around and say, No. The AI will say, Well, okay, but this is our phone number, our email address. If there's anything we can help you with in the future, please get in touch. No human intervention. It's all done very quickly, by SMS. Within a minute, you've either eliminated that lead so it can be wiped from the database, or you can follow it up and it's a lead you wouldn't have been able to get to before. We're very strict. We don't bug people if they say, No, that's it. That's fine with this. There's no point in keep on following people up with stuff that they're not interested in. The idea is to very quickly eliminate people we don't want to follow those Yeah, yeah.

Speaker 1  23:22  
So it's kind of like a James Bond. You just go around eliminating people,

Unknown Speaker  23:26  
yeah, something like that.

Speaker 1  23:29  
How do you balance the efficiency then of AI assisted copywriting, um, with maintaining a new a unique brand voice for your clients and the human touch. How do you do that when you know things like GPT, the latest GPT one, when that goes you know, like with the ability of Claude, to be able to put you know English together, that'll be stunning, absolutely.

Speaker 2  23:54  
But as you said earlier, nothing replaces your I don't know what you'd call it, just your instinct as to how the tone of voice, the brand, all comes together, and what you'd like to put off, I tend to always use AI as a starting point. It's never my final product, not ever, because you do have to infuse the brand personality into everything, and you need to use your human editing skills to refine and personalize everything, both for you and the client. AI is just a tool, isn't it? Final

Speaker 1  24:34  
question, what advice would you give to business leaders who are hesitant to invest in AI due to concerns about ROI or implementation challenges.

Speaker 2  24:44  
I think the longer you wait, the worst it will be for you in terms of the amount of money you'll have to spend to catch up, the amount of time you'll have to spend catching up. And I don't understand anyone. Hesitating to get involved in this, because it's just such an important development. You've only got to look at the figures of the investments are in the hundreds of billions of dollars, and the hardware that they're setting up to cope with all this. It's huge. It isn't going away. So I think that I'd recommend starting with small pilot projects to test the AI capabilities of both your staff and your company and all the facilities that you have, and look for things that you can do with the AI to actually improve your operations. Really, I think that would be the thing. Sit down with your team and set clear and measurable objectives. All the usual things that you would do, but make sure that you monitor them. Get a cross functional team to oversee the implementation. So you'll need somebody from compliance, somebody from logistics, somebody from accounts, somebody from customer service, and just talk to each other about how you can do this. The long term competitive advantages will be amazing if you get involved now, if you leave it too late, other people will jump across you and work out things that you could have worked out. Do you also got to talk to other companies, or even your competitors, to see how they're using this and suggest resources and partnerships to actually be able to enhance the way in which your team are looking at this, because we're all at different levels, and you know how it goes. In a brainstorming session, somebody Junior will come up with an idea which will blow everyone away because they're too involved in the nitty gritty of the way in which the AI hangs together and they miss the obvious. So I go back to our thing about communicating, talk to everyone about these things, to learn different stuff, and the way it all hangs together.

Speaker 1  27:16  
Well, that then brings me to a thought about creative or creativity, and also you don't want a silo. You don't want in the organization. It's got to be cross functional teams, because if you just have it run by legal or HR or accountants, it's just becomes a, you know, the loudest person or the highest paid person in the room's opinion, and everyone has to go with it. It's got to be cross functional. And you've got to allow those junior voices, as you talked about, you've got to allow those young kids voices to come into the mix and be taken seriously. Yeah,

Speaker 2  27:56  
I agree. The AI gods have obviously heard you, because today, perplexity released spaces, and spaces is designed to be exactly that. So you can set up a project within a company or a group of people, and everyone can contribute. And it's, it's a space on perplexity where you can refer back and forth to all of the research that everybody's done, people can make notes on it, get the AI to churn something and then put it up for everyone else to read very clever spaces. I haven't had a chance to look at it properly, but it, they released it today. Just

Speaker 1  28:35  
the just the sheer number of man hours that will will save, because all that data is in one repository which can be chewed over, you know, zillion times a second. And you know, everyone is contributing. That's brilliant. Gary, this has been an absolutely fantastic conversation, and brilliant. Thank you so much for for doing it at six o'clock in the evening, your time and three o'clock in the morning. My time is

Speaker 2  29:05  
Really Oh, I'm so sorry.

Lee: Really, it's fine.