Ken Roberts - AGI changing the entire advertising industry and popular belief systems

Speaker A: Hello, um, bonjour M. Nihau Comestas. Welcome to Champagne Strategy. Listen to this episode if you dare, but you've been warned, there's no going back. Thanks again for chatting to me. Um, but last time, uh, there was a lot of people who agree with you and sent me really good things like, oh, I love that you're exposing this. You know, it's been going for ages and then there's another 50% who are like get wrong, blah blah, blah, blah. And I had some really good feedback.

Speaker B: So couldn't it be 60, 40?

Speaker A: Yeah, I think it was actually, to be honest, because when people complain about things, it's generally if they're happy, they don't really DM you. Right. Unless they're super happy. So I just want to run through some of these and see what you have to say about some of these. So, uh, the first they said emotional creative doesn't get remembered. Is that true?

Speaker B: That's completely at odds with science. Our journey is to go from the marketing communications to the long term memory.

Speaker A: Mhm.

Speaker B: That's the journey. And what facilitates that? We understand that coding into the long term memory is aided by emotion. If I ask you to tell me about your earliest memories, you might have a memory of being in a cot, you maybe have a memory of your

Speaker A: first day at school or smell or

Speaker B: whatever, but I bet you there's an emotion associated with it. And by the way, one of the issues is that we cannot act without emotion. So our emotion is driving our behaviour constantly. We're not aware of it. So Dimasio estimates that we're aware of less than 5% of our emotion. So 95% of emotion is going on running in the background all the time, driving our behaviour and work has shown that the cognitive can be, and this is quite astounding, but it can be up to three seconds. The cognitive, which is, which is the rational. Why am I behaving the way I'm behaving can be up to three seconds behind the emotional decision that you've made. So it's just at odds with um, actually the neuroscience. And so John, the folks that express those opinions, you'll almost always discover that they haven't done first principles, haven't understood the first principles of how it's black or white. Yeah, the brain works and I suppose

Speaker A: that segues really nicely into this idea that we can somehow pretest creative that has a um, bigger emotional effect on people there before it goes to market. And that can be predictive of market share from that creative. So Some people say that these pre testing tools can create creative, um, filter out the bad stuff and make sure that we have the best chance of getting through and creating an emotional response. What's been your experience with pre testing?

Speaker B: It is such a massive topic.

Speaker A: M. This, uh, came from a pre testing company's representative.

Speaker B: Right. Well, the great benefit of my years in the category has been on my Lord, I've seen this stuff coming and going and in its different iterations, different forms. 2011, 2013 was the high watermark for emotion because folks were beginning to understand how to read it. We had people at conferences talking about that they had measuring people's emotion by them wearing T shirts and that the T shirt would pick up a uh, physiological response, like a galvanic skin response, you know, like their temperature rise, you know. That's true. There is a physiological tell for emotion. But how did that help advertising?

Speaker A: You sell something to create market share. But that's the next question.

Speaker B: So then most of the organizations that are attempting to measure emotion are uh, using Darwinian, and I'm not exaggerating, they literally are using Darwinian ways of, you know, in terms of what are the primary emotions?

Speaker A: This is Ekman's sort of base theory.

Speaker B: Exactly right. Ekman. One of his emotions was discussed. So if you look at um, you know, the classic popular measures of emotion, like you know, showing people photos and asking which of these photos best represents how you feel and, and one of them is disgust. Well, uh, all of our primary research which has been extensive, we've been published in the, in the top tier academic journals, no one else has, uh, in this field. Um, that's a marketing research firm. That is disgust is not an emotion associated with behaviour, but pride is. And they don't measure pride. So as marketers we understand about the pride of a new car, the pride of the house. So pretest. It's really tricky here by the way. If you think emotion is associated with some big event, you know, tears streaming down someone's cheek, that's just not right. I can get an emotion from seeing a price. I can see a really good price, a bargain. And that could create happiness. It can.

Speaker A: Excitement.

Speaker B: Yeah, whatever. So what we contend is that for every marketing offer, there should be a triple play, There should be an elicitation of an emotion, there should be a price. And it doesn't need to be explicit, the price.

Speaker A: Okay, so we've got some kind of value.

Speaker B: No price as in pure price. But we have um, lots and lots of clients that like, we have clients that have big franchise groups and the franchisees, you know, so we will have a client that will have a, uh, franchise in Neenah, Wisconsin and they'll have another franchise in Times Square, New York. And they can't set a consistent price. So they can't go on linear TV and say a price. But they need to be able to communicate price effectiveness. And so the volume of food that's on the plate, the talent that they're using, they're all non price queues for saying we're price competitive.

Speaker A: So it doesn't have to be a uh, figure numeral, but a figure definitely

Speaker B: works best, I can assure you.

Speaker A: But that's the most expensive, explicit of the price.

Speaker B: Yeah, yeah, but it can be subtle and then there's got to be a quality driver. And I, I, I hope we, we talk a little bit about um, Starbucks, because that's a classic that I, that I would love to touch on. But, but in your, you have to

Speaker A: come back to that. And you reminded me of another thing, um, around these, these three things. We'll get to it. So, okay, apparently system one does measure pride.

Speaker B: Do they?

Speaker A: So.

Speaker B: Right. They've added it, I take.

Speaker A: That's what I was told.

Speaker B: Thanks, John. Karen.

Speaker A: Phil Barton says Damasio talked about Dimasio last time as the basis of a lot of that.

Speaker B: I've got his book, you know.

Speaker A: Yeah, well, some people say. Okay, I must read it. Apparently Dimasio has been debunked. So uh, you know the one where he's gambling, apparently, uh.

Speaker B: Oh, that's the Iowa. The Iowa gambling experiment is not Dimasios. It predates Dimasio.

Speaker A: Oh, okay. So, okay. But that apparently has been debunked. Okay, so let's just skip that.

Speaker B: I don't know if that has been debunked, has it?

Speaker A: Apparently it wasn't replicated the same way.

Speaker B: Oh, I've seen it replicated.

Speaker A: Okay. Okay.

Speaker B: Yeah, I've seen it replicated in academic journals.

Speaker A: Okay, good.

Speaker B: Yeah. But what people are. So everyone knows the Iowa, uh, gambling task. I haven't looked at it for a long time, but basically what it is is that they're measuring physiological emotion and they're saying that the person shows a physiological tell before they start behaving. So one deck of cards is stacked and so as the cards keep coming out, the person's non conscious has already picked up that one deck is stacked before they start actually behaving. And unless I'm mistaken, that's what that finding was. And I've seen it. It's Been replicated. So I'm surprised that Phil says that. But Phil does know he's material.

Speaker A: This is the quote, caution against the idea that emotional based signals decide for us other than in extreme situations. So maybe there's an extreme situation.

Speaker B: Yeah, well, we have a triple play and the triple play is that there is a price, there is quality and there is emotion. So we would agree with him. It's not purely emotional. And his point is really good because you've seen organization after organization after organization that have tried the emotional anthem. They've done the big emotional out there and it's had no result. We could talk about.

Speaker A: I see that all the time.

Speaker B: We, we see it all the time.

Speaker A: I call it the emotional puff piece that the agency has sold, wins awards

Speaker B: and it has zero effect. And they spin this story about it's the long and you're just not patient enough.

Speaker A: Yes, yes.

Speaker B: And you know what? The long never comes.

Speaker A: I see that all the time actually.

Speaker B: So do we.

Speaker A: I always say like uh, how long? And they can never answer that question.

Speaker B: No, as long as the. Well, they know the CMO is not going to last that long.

Speaker A: Well, apparently I was looking at this research, um, the Half Life advertising. There's a couple of studies, but I looked at like four of them. And the Half Life is about four weeks of um, this sort of lag thing. So you get half the value response or market share or sales in four weeks and then the other half can extend from 12 to 24 months. But very rare to go beyond that. And obviously it trails off, right?

Speaker B: Yeah.

Speaker A: And obviously it really depends on the creative as well company and other macroeconomic conditions. So. But uh, again there is a half life and there is some sort of like curve that's kind of consistent between these four. But I wouldn't say that goes any more than 24 months.

Speaker B: We've seen creator that, as you say, I mean it's just so variable. But at the very far extreme we have seen creative that keeps. It's still working at three years after the first. Yes, after the first area.

Speaker A: Wow. It sticks in the brain so well.

Speaker B: Uh, it's just reinforcing and it's so well steeped in foundation in the foundation of the drivers of the category that it just keeps reminding people that's why

Speaker A: they're going there, that's why they're without other communication. Reinforcing or rejigging the memory.

Speaker B: No, there's always. And everything else. But yeah, there's always a. Yeah, yeah.

Speaker A: But okay.

Speaker B: Very variable.

Speaker A: Tom Barton says system One and two. Okay, so we're talking about Kahneman System one and two. Sort of thinking the book, you know, um, you either do one or the other. And he kind of says it's a bit more nuanced. Like it's sort of, it's a bit of both of ones and sort of ebbs and flows depending on the context. Like I think there's some people that believe that it's like you're either emotional or you're rational. Like. So it's more nuanced than that.

Speaker B: Totally, yeah. Okay, so uh, you mentioned system one. The difference between us and them and many differences, but the main difference is that we can tell you to uh, the second decimal place, the relative importance of the emotion and the decision. So prepaid mobile phones, prepaid cell phones going into 711 and doing a top up. It's 5% emotional and 95% rational. First time mums buying their diapers for the first time. It's 70% emotional and only 30% rational. And then the 30% rational is just one rational drive, which is leakage protection. So it depends on the product, it depends on the buyer. You can't make those kind of statements perfect.

Speaker A: And then the last one, uh, so Shannon Bossard is a neuroscientist. I think this came out of our last discussion. I'm like, there's a difference between, um, creating pieces of creative to be emotional and then the person, whether they view that or absorb that, whether they get an emotional response and then whether that actually leads to as something else. So he said there's a difference between affective information, uh, so something that tends to affect emotion and there's the emotional response itself. And then there's the feeling that's created, uh, which is a perception of the emotion which is sometimes, like you said, conscious, uh, or unconscious. Yeah. Is that a good way to think about it as opposed to just go, hey, it's either emotional or

Speaker B: he's absolutely right in characterizing a process.

Speaker A: Yes.

Speaker B: Think for a moment that you're driving a car and a, uh, person on a bike loses balance, A door opens, they come out in front of your car, you quickly sway to miss them. You miss them, you pull over and the adrenaline is running through your body and you feel, you know, you feel your heart racing, you're actually shaking and you feel that emotion. So that is an emotion made conscious. So when those folks say very few emotions ever get to a conscious process, but it's interesting to understand. Okay, that's one extreme. What about more modest emotions?

Speaker A: Like, just say Burger King has a, uh, special, like, snack menu for $2. 55.

Speaker B: That could be a little modest. Overly modest Burger King, but. No, but let's go with that. So there is a progression. So there's an emotional response. Then there is a feeling. So we go from emotion to feeling. A feeling is an emotion colored in by memory. A feeling is an emotion colored in by memory. So I'm walking down a track, I hear a rustle of a snake. I get the emotional response is ahead of my cognitive response. I'm reeling, I'm looking down. And now my memory is coming in to, what is that? It's probably a snake. So now I'm in a feeling stage. And so dramatic is that response that it's now progressed to a cognition. To cognition. So now you've gone from emotion to feeling to cognition.

Speaker A: Okay, so you're conscious of that feeling.

Speaker B: Exactly.

Speaker A: That emotional state.

Speaker B: Exactly.

Speaker A: Okay, interesting. So I think the key thing here is that you can create a piece of creative. We're talking about marketing communications context, or a video or an ad or whatever. You can say that this is effective information, like, or effective creative, intended to elicit, uh, a response. But whether it does or not is a, a pretty big gap. Correct?

Speaker B: Well, we understand in our methodology, we do a calibration of how you're feeling before we show you any stimulus.

Speaker A: Oh, like a baseline?

Speaker B: Yeah.

Speaker A: I think that's pretty important, isn't it?

Speaker B: Yeah, totally important.

Speaker A: I think a lot of other testing,

Speaker B: no one else does it that I'm aware of or maybe, uh, no, I think I do know someone else that does it. So I shouldn't say that, but. So find out your base, then show you the stimulus. We can see a change in your response at the emotional level and at the cognitive level. And we can see at the cognitive level which drivers have changed. So we can see whether it's improved your price perception or whether it's a quality. And what's the quality driver.

Speaker A: So some people asked, how does that work in your process? What can you talk about the process? Like, do you put electrodes in people?

Speaker B: No. No, no.

Speaker A: Eye tracking or.

Speaker B: No. So CT scanning. Oh, look, this is the, uh, this

Speaker A: is the whole thing because people ask, like, what's the difference between your testing method and other ones? Like, tell me more about Ken's method and for method.

Speaker B: So we talked before about going back to 2011, 2013, all that period where there was charlatans telling all sorts of stories about how, um, you could measure emotion. You know, put on a Hat put on electrodes, you know, like, for example, you know, how deep do those electrodes go in terms of measuring response? Um, you know, galvanic skin response, you know, wearing T shirts, all of that kind of stuff. It's all been debunked. It's been totally debunked.

Speaker A: Well, why is that? Is it not replicable?

Speaker B: Because it's just measuring. Well, the most important thing is that it's not measuring discrete emotions, anything that's obvious. Yeah, well, it's just measuring a physiological response. Uh, you know, so they say, oh, right brain, left brain, what emotion is it pride? Is it happiness? Is it anxiety? You don't even know if you're not measuring discrete emotions. And it's. The marketing fraternity want to understand discrete emotions. So, for example, in mortgages, the discrete emotion that drives consumption behaviour for mortgages is love. And it's not love for the bank. If you show that to the bank, the bank will go, yes, they do love us deep down. But of course, it's about the person bidding, thinking about their buying another room for their kids or whatever. So love is that driver. But how do we measure it? We measure it implicitly. So implicit measurement, uh, is about how many milliseconds does someone register this response before we know that it's no longer emotional and it's probably cognitive. So in our case, um, this is a little technical, but in our case, the modal class is under 1 second for response to our survey, to our implicit measurement of emotion. Under one second. Now, the modal class for a cognitive question is. Is three seconds. Three seconds and more. So you're getting a response in a period that is so quick that we say minimum cognitive interference.

Speaker A: Okay. Okay. Because the cognition could override that once they're conscious of it. Yeah, yeah.

Speaker B: Well, they'll post, rationalise. You know, it's like focus groups are the worst. You know, you get, how does that make you feel? And all you think about is how much you hate the person next to you.

Speaker A: Yeah. Or you don't know how you feel and you're like, well. And then you make something up. Yeah.

Speaker B: Right. And we don't have access to how we feel. So.

Speaker A: So that time, that time, um, sliver of measurement is really critical.

Speaker B: Really critical.

Speaker A: Okay. Okay.

Speaker B: Yeah. So we, um, we measure that we know. And so just as, um, in survey design, you'll find they talk about speeders. So they look at people that go through surveys really quickly and they know that they're not actually reading the question and answering. They're just in it for the incentive. And so when we get that data out, they think they might have gamed the system. But we actually don't include them in the survey. But we also don't include uh, in the emotion people that take that ruminate the people that take too long to decide what their emotional response is to this. So system one shows a series of photos and then it has variability of the photos. We show an animation of a metaphor. So we have nine metaphors and so they're fully gamified. So you go to the next one, it gives you a quick review of the full animation, uh, of the full 0 to 10. So beneath is a 0 to 10 scale and then it stops and you move it. So from the time you move it to the time you take your finger off it, that's got to be under one second.

Speaker A: Okay.

Speaker B: So that's our measure of, that's our implicit measure of emotion. We really strongly believe we're the best in the world at doing it.

Speaker A: Great. Okay.

Speaker B: Yeah. And it's patented.

Speaker A: Yeah. Because I've seen um, I've seen some people do some tests on these because you know a lot of these micro research, um, companies that do this testing uh, use panel providers.

Speaker B: Right.

Speaker A: M And uh, this guy um, could control the environment. So he kind of white label the panel provider inside his own web page and then he put um, that sort of hot jar sort of tracking on it. And he figured out a lot of them are just um, using auto clickers or just fake. So there's like a huge percentage. And this is a big data provider, panel provider. And he figured out that like a huge percentage of those are uh, fake. So therefore you know, the majority of your responsibly fake.

Speaker B: Yeah, look, we've had examples like that but um, particularly in other countries, we uh, haven't had those examples in Australia. They're extraordinarily insulting because it implies that when the data science folks get the data that they don't look at anything, they just give you a mean and a frequency. And that's the answer. And that's just not actually how the good quality research firms work.

Speaker A: You remind me of something, uh, what you said there. So there's this other lady, um, how emotions are made. Uh, her name is Lisa Feldman Barrett. But this is an interesting quote. She said now we see that this rough and fast system is actually remarkably accurate. Even the fast system has access to logical reasoning, meaning intuition might be correct more often than we'd expect by chance. And then she has this other quote even after A century of effort. This is what reminded me about emotional measurement, like, not being easy. Uh, she said, after a century of effort, scientific research has not revealed a consistent physical fingerprint for even a single emotion.

Speaker B: A consistent fingerprint, yeah.

Speaker A: And then, uh, when scientists attach electrodes to a person's face, you know, we're talking about that, the physiological response and muscles and how they actually move during the experience or emotion, they find tremendous variety, not euphoria.

Speaker B: Oh, totally. No, that's. That's completely right. Yeah, that. That's like when we first started looking at this field, people were saying that there was a, ah, a love. And there was a buy, like we'd

Speaker A: find or something we discard.

Speaker B: We've discovered the buy button.

Speaker A: It's right here.

Speaker B: Yeah, yeah. We just have to. And then.

Speaker A: But that's fantasy. You're saying, like, everyone's so variable. Yeah, yeah, yeah.

Speaker B: And, um, very good quality. Scholars have worked that out.

Speaker A: Okay.

Speaker B: It's a variation. And it's really important because it goes to us as brand builders to say, you know, what we're attempting to do is create this neural pathway between behavior and an emotion associated with that brand.

Speaker A: Yes, yes. Um, the other thing was, uh, I know Byron Sharp doesn't believe in this, but this, um. Ah, in advertising, sometimes they say there's a left brain thinking, a right brain thinking. Not to believe in that. Right. It's rational or something completely debunked as well.

Speaker B: Yeah, that's right.

Speaker A: Why is that so? Why do so many people believe that?

Speaker B: Is it just one of those, uh. We mentioned System one on. Want to give him too much air. But I was at a conference. I was speaking after John Kerran one time, and it was a big conference. And he said, uh. He walked out and he held out Kahneman's book.

Speaker A: Okay.

Speaker B: And he said, hands up who's got thinking fast and slow? And practically every hand went up. And he said, leave your hand up, uh, if you've read it or if you've finished reading it. And I reckon there was like three hands in the whole room. And he said, oh, there's a ten pound note at the back of it. You know, you might want to go and finish it.

Speaker A: It's a heavy book. Yeah, it's the same as Halborn Screw like, oh, I quoted this Byron shop thing. Have you actually read every page? Like. No.

Speaker B: Well, this is. I mean, Andrew Ehrenberg has been dead a long time.

Speaker A: Yes.

Speaker B: And certainly he missed the whole thing about emotion. So, you know, thanks, Ehrenberg Bass, for bringing forward his Work. Amazing. Really important. But there is, you know, there is an ongoing understanding of how markets work.

Speaker A: Yes.

Speaker B: And, um, that's where Ritson is so important because people aren't reading the literature and they never will. Whereas Ritson is reading the literature and he's trying to curate a, uh, basic truth about marketing. And as you and I have spoken about before, he's basically replaced Philip Kotler. Philip kotler in the 70s, 80s and 90s, had the Encyclopedia of Marketing. You know, it was inaccessible as well in terms of, you know, a book. Whereas you would have to say that Ritson has been singularly responsible for bringing the marketing fraternity up to here. Uh, my sense is they should be here, but as a consequence of what they've learned from him, they are on a journey.

Speaker A: Yeah. I think there's a big difference between, like, how things actually work, which is not always glamorous, or, you know, you can present on stage and get a really good response as a talk. And then it's like how people think they work. Yeah. And it's like sometimes that gap is the problem.

Speaker B: But so I'm going to do.

Speaker A: Ken, I'm, um, really interested in what you're doing because I've heard you work with a couple of clients, um, a lot of which you can't disclose. Ah. But I think m with your new company, brandcomms AI, you're really using Gen AI in this video. Put your system into something that's a bit easy to produce for. For companies. Can you tell me about that and how it works and what's happening?

Speaker B: I've been on a few panels lately. I was on a panel with, uh, Russell Howcroft recently, and he said what proportion. His question was, what proportion of advertising will be replaced by Gen AI by 2030? And my answer is 100%. It's truly troubling for the whole category to think about it, if you believe it. And there's been studies shown, uh, that undertaken that show that creativity is under the greatest threat of gen AI, uh, in terms of being 100% replaced.

Speaker A: Well, actually, I was doing a tour of an office of a very large gambling firm, and they said that not one single piece of creative they send out now and they do hundreds a day, um, doesn't use AI. So there's all some kind of element where there's a super or it's an image or video. So they're like, everything we use every day is AI. And it's becoming more and more and more because we didn't need as many people.

Speaker B: Yeah.

Speaker A: So I believe it.

Speaker B: Yeah. The problem with gen AI and advertising is that what it's mostly going to do is deliver bad uh, creative faster. That's what it's primarily going to do. If all you do is take genai and agentic AI and replicate the practices of the past, that's all you'll get bad creative faster. So you'll still have half. Ah, my creatives not working. Uh, problem. That's what forethought is doing and that's why we've spent millions of dollars developing brandcomms AI. And what that's doing is that we've created a content store which is based on actual behavioral data. And so uh, we're creating advertising that is guaranteed and that's what we guarantee that it will drive behaviour, behavior meaning outcomes in market.

Speaker A: Great, okay. So you know, purchase, trial, that kind of thing as opposed to just making a motion and then go okay, I'll buy someone else. Yeah. You're saying you can link it to behavior.

Speaker B: Okay, absolutely.

Speaker A: Because that's a holy grail really of advertising.

Speaker B: Yeah. I'd say we've got lots of clients that have discovered it but okay.

Speaker A: Does this stuff win awards at Cannes?

Speaker B: And uh, I don't know of any of our clients that care.

Speaker A: Isn't that interesting though because there's that whole glamorous side of award winning agency creative um, that gets lauded and talked about and industry press.

Speaker B: The creative side.

Speaker A: Yeah. And then there's stuff that actually works.

Speaker B: Yeah. But then there's the, then there's the you know, like I'm not getting rid of my creative agency because they're the cool guys and they take me to Cannes. Yeah. Like the economics of ah, generative AI and agentic AI will weed out that entire sect of people. I uh, wouldn't mind talking in a moment about the industry. So I think we're in really, really a bad place for the industry. Uh, they've been forced down in terms of pricing. The advertising agencies, they're not paying their staff, uh enough money cutting staff. They're creating staff turnover where clients are getting juniors and constant turnover. The holding companies emerging and losing money and shrinking pretty quickly. And creative is being brought in house. So you know, it's kind of like the writing is on the wall. The writing is on the wall just like it was on the wall for Blockbuster. Just like it was on the wall for Kodak and just like it was on the wall for everyone that was in the horse and buggy industry. And you ask yourself, but it was surely obvious where we were going with the horse and buggy. You know, cars went four times faster than horses. You didn't have to have a stable. It was only 10% of the cost. So surely everyone that was in the horse and buggy industry said, writing's on the wall, let's get over and start making cars. And yet you know that the only company in the world that transitioned from that old industry to the new industry was Holden in Australia. M. And they went from making saddles in South Australia to making cars. But other than that, every other company

Speaker A: that was making horse equipment.

Speaker B: Yeah, whatever, all went out of business. And you could say, what, didn't they see it coming? Like the horse and buggy lobby had, uh, churches ring their bell every time a pedestrian was killed. Because they were saying, you know, this is going to kill people. Look how fast they're going.

Speaker A: It sounds like Tesla, you know, all these companies against Tesla, Tesla, electric car companies. Yeah.

Speaker B: But it sounds like the advertising industry that you go, surely they saw it coming. So the analogy that I think about all the time is that you're in the ocean, you're a surfer, and you're watching out. You don't look at the shore, you're looking at the waves, and you see this wave coming, and Lord almighty, it's the biggest wave you've seen ever. And you're having to make a decision, am I going to dive under it or am I going to start paddling and catch it? And if you dive under it and then come out and decide, I should have caught that, then you can swim as hard as you like. You'll probably never catch it. But if you want to catch it, you've got to paddle hard. You've got to paddle hard to get on top of that wave. And that's where we're at. So I gave that analogy at a conference at the ARF in New York a couple of months ago. Uh, they had me as the zealot. I was the.

Speaker A: That's what I'm brought in for sometimes. Like a dissenter.

Speaker B: Yeah, I'm the Genai zealot.

Speaker A: Okay.

Speaker B: And it was like thousands of advertising people I was talking to. But m. Honestly, I think that if those folks are going to transition, they need to start thinking about transitioning right now. Because I still believe what I said to, uh, Russell. And that is, I don't think that there'll be. I really believe there'll be no old creative industry in. By, you know, in five years. It's. It's fast.

Speaker A: So maybe you'll make things like this redundant. Like Gen AAA would be so indistinguishable. Like, you know.

Speaker B: Oh, totally.

Speaker A: Like VO3. We're recording. We're actually physically recording now. But I want to talk about this because how do we know that something is not fake? Do we care? This counterintuitive, like flight back to authentic things that can't be faked, like in person events and things like that.

Speaker B: Or we are adding flaws to people's complexion in Gen AI at the moment

Speaker A: to make it look more impersonal.

Speaker B: To make it look more person. Yeah, make it look more authentic. In the space of less than eight months, we have gone from being able to detect to not being able to detect Gen AI.

Speaker A: Wow.

Speaker B: And. And we're only at the beginning of the journey.

Speaker A: Yeah.

Speaker B: So lip sync and things like that. I mean, they're all. They're all happening as a. Like the guys here have. One of the big challenges has been where there's a character that's repeated throughout the ad and generating that same character. Things like that.

Speaker A: Like an avatar or something like that?

Speaker B: No, no, person.

Speaker A: Oh, okay.

Speaker B: So you've created a person at the beginning of the ad, then you want to show them consuming the product. Then you want to. So those are the kinds of things that they're mastering at the moment.

Speaker A: So actors are basically totally, what, sound generation, filming, cinematography.

Speaker B: The way that AgentIQ is working is that. Is that we are creating these autonomous agents that replicate every job in a creative agency.

Speaker A: Okay. Specialized roles of each one of them.

Speaker B: So there's an agent that's the creative director, there's an agent that's responsible for scenery. There's an agent. And so all those agents are learning and they are, uh, acting autonomously to create ads. So it's not enough to just be creating the ad. The ad has to be actually driving behavior.

Speaker A: Yes, exactly.

Speaker B: And so it has to go to a content store that teaches it what causes behavior to occur in this category. And that's what brandcomms AI does.

Speaker A: Okay. Can you talk about any of the work you're doing?

Speaker B: Sure. It's all over. It's. So we have master service agreements with every client, and every client says that we're not allowed to say that we're even working for them. So I'm gonna have to talk in generality.

Speaker A: Sure.

Speaker B: So one of the largest confectionary businesses in the world is using it. One of the large brewers in the world is using it. Organizations that are spending over a billion dollars a year on advertising are using it, uh, for that one Brand are

Speaker A: you using it like tinkering with it or are they using to replace everything?

Speaker B: No, there's three cases, there's three use cases. Use case number one is they're using it to develop um, great creative concepts. So they're taking the animatic and giving it to the creative agency.

Speaker A: So it's kind of like the screenplay or storybook you did before they're doing for that and give it to a physical production team or. That's right.

Speaker B: And that's use case one. And that's been fairly common at this stage. There's a reluctance of brands to be uh, seen to be doing gen AI. So they're using gen AI to create the spot, you know, 32nd, 6 second or whatever and then they're handing it to a production company or to their creative. Now I've got to say that in most cases that's been okay, but in a lot, not some cases it's been okay. In most cases it's caused the quality of the spot to go backwards. So what we pre tested from brandcomms AI to what we pretested from what they produced was not as good.

Speaker A: So they stuffed it up, they did

Speaker B: stuff it up but there was a marginal reduction. And by the way, one of the things we have noticed is the creative agencies that have taken that work have

Speaker A: not understood the fundamentals.

Speaker B: No, they haven't listed us as uh, part of the process, but that's another story. So that's use case one and in m our view that use case will go. Yeah, that's where we're learning how to use Gen AI and agentic AI in advertising. Use case number two is I think the biggest area we're seeing which is the creative has been brought in house. At least part of it has been brought in house. And we want tools that are going to enable us to ah, outperform the stuff that's not been brought in house.

Speaker A: Correct.

Speaker B: So we've got quite a few clients that have got uh, certain brands or certain geographies that are being done in house and still have a relationship with the creative agency. And those guys are using us a fair bit. There's a third use case which is organizations that simply can't. They don't have the budget to get to every geography or for every product. We call this the long tail of neglect. And Genai enables them to produce uh, super high quality spots to put into markets that have been neglected.

Speaker A: Just like mass personalization of a core concept.

Speaker B: That's where we're going. Okay, that's where we're going.

Speaker A: So a lot of franchises will obviously like that because the pricing differs.

Speaker B: Absolutely, yeah. By geography. Um, but yeah, and that's exactly what

Speaker A: the point is as well.

Speaker B: So the algorithm that is creating the ad, um, based on the content store, increasingly that will be creating the marketing communication for John. Increasingly. In fact, we think by 2030 gen AI and agentic AI, more rightly agentic AI will be creating individual ads, individual marketing communications. And you could say it's doing that now. We're getting that now. Um, Adobe and organizations like that would say that they're serving up individual ads now, but what they're doing is that they're serving them up to segments. They're not serving them up to John. Or they might be, but they haven't got the category behavioural data.

Speaker A: So we haven't got to the um, Blade Runner sort of style personalized ash.

Speaker B: No, not yet. No.

Speaker A: But that's coming.

Speaker B: It's coming.

Speaker A: Yeah.

Speaker B: I no doubt that it's coming now.

Speaker A: Um, you did send me. I mean, I kind of agree with. I see it already. I think there's a lot of people, obviously if they spend 20 years in their career, they're going to have a bit of, uh, friction around the reality of the situation. So I get why they're probably saying things against it. Um, but, uh, the Starbucks thing, so really interesting. You sent me a report on Starbucks. Yes, obviously. The context is, um, it was a pretty successful company. They started raising the prices a lot. They got another CEO. Was it ex McKinsey, I think to come in?

Speaker B: No.

Speaker A: Or was it X Bain or.

Speaker B: No, no, no, no, he was, uh, he was. He's from the category.

Speaker A: From the category. Okay. Uh, then he was the CEO for a little while and then the share price started to cradle.

Speaker B: Oh, no, you're right. Okay. The previous. The pre one. Yeah, no, you're damn right.

Speaker A: Okay. So I think it was McKinsey and then he.

Speaker B: I don't know where he came from,

Speaker A: but I think it was one of those consultancies anyway. Ah. Then that created went downhill and then they've got to this crisis point of loss of shareholder value where they've got a new CEO and I think it was X. Uh, another sort of food.

Speaker B: That's right.

Speaker A: I forget they won, but big food franchise Chipotle. Uh, that's right, it was. And he's brought in to obviously recover the situation and put that advertising, um, creative out. Um, and I saw the one that you sent me which was, um, you know, writing the name on the app, I think it wasn't it. Which is a new campaign. And you're a bit skeptical about whether that would work.

Speaker B: I wasn't skeptical.

Speaker A: Okay.

Speaker B: No, we just outright said it won't work. No, no, no. So this is your classic example of category drivers. So you've got the new CEO, uh, he brings the CMO from Chipotle in as well. And the first thing she does is that she terminates the agency. The agency had only been there three months, and, uh, that was wpp and she replaced them with Anomaly. And, uh, so they came up with a creative. And the creative was, uh, nostalgia. Like, you know, do you remember how good it was back in the day? Starbucks. So. And one of the things about nostalgia, but the thing that they remembered about the third place, that it had WI fi, it had chairs, you could sit down, you could order a coffee, you could do a bit of work. It was, it was a great vibe. And then there was also personalization. People would write, you know, john on your cup.

Speaker A: And then they'd say, hey, John, your coffee's ready.

Speaker B: Yeah, and here's your cup and there's John. And maybe they didn't know your name and they wrote you a note, you know. Yeah, get a shave, John.

Speaker A: You know, love heart, love the stubble.

Speaker B: Um, so m. That's your classic creative coming in. We've got this idea, whereas we are in the category. So we were getting to the stage where we're past the stage where no client will show our work in terms of where it came from and stuff like that. So we said we're going to have to go and do a work for a client, do work on a client that is not a client of ours. So we did Starbuck because we had a content store because we've doing their competitor.

Speaker A: Okay, so you have a client that's in the same category. That's right. But, uh, you also know the category so well that you know that there's idea is probably flawed.

Speaker B: We know that it's not category drivers. Nostalgia is not a category driver. Third place is not a category driver. Personalization is not a category driver. What we know is there's a triple play, and the triple play is happiness. Uh, that it's, uh, about, um, price. Of course, there's always a price element. And it's about rewards. So the loyalty program, the rewards. But the biggest driver is the ability to customize my order. Uh, the ability to customize my order. So I want this, this and this. Right? And happiness. So the creative that brand cons took out of the content store created that. So it did the. So in these limited time offers or these kinds of fast moving areas we often have a cultural overlay to make it contemporary. So the cultural overlay that was chosen was Bridgerton.

Speaker A: Bridgerton.

Speaker B: Bridgerton. It's a, it's a television series.

Speaker A: Oh that's the uk US actually.

Speaker B: Yeah, yeah, damn right. It was very popular at the time.

Speaker A: Okay so this periodic UK what was it like 19th century. Yeah, 18th century thing.

Speaker B: So, so in, in our post production of us talking, we'll put that in there, we'll show them what the ad looks like. We'll show them what the anomaly ad looks like. Why would you pick that brand? Comms wanted to be contemporary and knew that that was what was contemporary. Yeah. So there's another twist to this and the twist is are you trying to sell additional services to existing customers? Are you trying to get the same customer to buy more or are you trying to get what's called in the category the NP3 the non past three months to come back.

Speaker A: Oh, uh, so like elapsed customer.

Speaker B: That's right, yeah.

Speaker A: Okay. Which the end would be very large, I'm guessing.

Speaker B: Very large.

Speaker A: Yeah.

Speaker B: So the Bridgerton spot with those three drivers was targeted at the non past three month guests. So when you look at the results from the pre test you see that it really smashes it out of the park. Statistically significantly lifts buying intention from non past three month guests. So that's marketing strategy 101.

Speaker A: Yeah. Diagnosed properly. Which obviously that in too.

Speaker B: Yeah. Well I know I don't want to, I don't know that. I don't know what was the stimulus for and I'm sure that, well sometimes

Speaker A: it's political thing like hey um, I'm a new CEO and make a statement, we're going back to the old way. Let's have a campaign that reflects that kind of. Yeah, maybe that's where it came from.

Speaker B: And um, they're in trouble. I mean it's contracting and he's, he's.

Speaker A: How are they going right now?

Speaker B: Like they're contracting.

Speaker A: Yeah.

Speaker B: And he's saying um, I'm fixing it. And by the way we would, we would say that It'll take him 18 months to fix it anyway. So you know, but that spot did not, did not even in our view did not even begin to fix it. And then if you look at the actual sales data from the period that it was on, it didn't fix it.

Speaker A: Yeah.

Speaker B: So it's kind of like how can you argue otherwise?

Speaker A: So It's a good test of the methodology, I suppose.

Speaker B: Look, the big issue at the moment, John, is the clients are scared of using Gen AI because of um, unforeseen legal issues or seen market issues. And so that's what's, that's. We've literally been working for a year with one client on, on the legal. And, and we're having clients approaching us, asking us for indemnities and stuff like that. If they. So we're, we're. This is the, this is the costs that the, the first firms will bear.

Speaker A: Yes.

Speaker B: And we're bearing them.

Speaker A: And I suppose small business don't care because there's not much to lose. But large corporate, that's a legal risk for them completely. Yeah, yeah, yeah. So people are dabbling with it on the side but not putting into production always. Or maybe not.

Speaker B: No, I think it's mainstream. It's on the edge of mainstream now. I would say by middle of 2026 it'll be utterly mainstream. It's not the stuff that Coca Cola did for Christmas.

Speaker A: That is because I saw they do this thing with Adobe and. Oh, the, the Gen AI Christmas thing.

Speaker B: Yeah, like that. That is, you know, even when they put that out there, it was infantile in terms of its quality standards.

Speaker A: I did see that.

Speaker B: And it's not because I'm now looking at it in terms of capability that we have now and judging it back on. No, back then there was far better quality that could have been done than what was done there. And I can show you stuff that we were doing at that time that was.

Speaker A: I do like give them some credit for like trying something as a big brand and going, hey, let's. I mean most brands would go no, no until it's polished. But they were just like, no, let's.

Speaker B: Yeah. Now my angle is not how brave they were and they to be commended for certain, but about people looking at that and criticizing it for what? You know.

Speaker A: But who? Advertising.

Speaker B: Yeah.

Speaker A: No one else cares.

Speaker B: No. Well, that's absolutely right. That's critical. And the echo chamber of uh, advertising folks talking about the shortcomings of Genai, they're constantly telling each other and they, you know what, you know what they're mimicking? They're mimicking Kodak, Blockbuster and the Saddlers. That's what they're mimicking and that's why they're looking at the beach instead of the wave.

Speaker A: Mhm. Okay, so where is this going then? Let's just say the dynamics of the industry were you had an agency they could Create creative. Um, there was a media buyer who knew how to buy the media, match it to up. Then there was a measurement system system on the back end. And like, you know, that's why you need an agency. So you're saying that in the future we don't really need that you might have an agency? I don't think so either. You might have an internal capability and you have strategy, maybe some consultants coming in doing in house strategy or even not, not having a need for it anymore because it's all contained in the system.

Speaker B: Where are we heading? I, uh, know I think that there will be, there'll be folks that will still be consulting and, and research. Yeah, I just think by 2030, uh, this will be part of the stack and that you will simply. It's generative AI and agentic AI. So it's not just large language models, but there will be large language models as your interface that you will just simply ask for a campaign. And by the way, the content stores that we're building for clients have got as input the brand tracker. So it's already writing, it's already writing the creative brief based on what it's seeing happening in market and it's already then creating the ad to address what it's seeing in market.

Speaker A: Uh, so it's taking the brand tracking results going, hey, you're falling apart in this. We need this creative to then solve that problem.

Speaker B: These are the category drivers. This is where you're underperforming. This is what you need to do. Here's the brief, here's the creative.

Speaker A: So almost the research diagnosis is becoming audible.

Speaker B: Way to describe it. Totally. Yeah, totally. So, um, so what does that mean then?

Speaker A: Then the media buyer is more important. The human attention is more important because the barest entry of creating video content now like almost going to zero.

Speaker B: I think that the, the media buyers, you know, when we look at the, the ecosystem of media and advertising, the media folks adopted quantitative measures two, two decades ago and the advertising industry still hasn't. So media is far more adept at actually this new environment. But um, uh, agents will do that job.

Speaker A: Yep, yep.

Speaker B: Yeah.

Speaker A: Rather than humans.

Speaker B: Yeah.

Speaker A: At a media agency going a spreadsheet and flighting.

Speaker B: Yeah.

Speaker A: Spots.

Speaker B: Yeah. And they'll optimize it.

Speaker A: I mean it already is in some channels like you know, Meta and Google systems. All rt.

Speaker B: Yeah. In a lot of instances what we're talking about has actually happened.

Speaker A: Yeah, exactly.

Speaker B: In some instances. But it would still be new news to a lot of people.

Speaker A: True, true, true.

Speaker B: Okay, well, so where's it going? I mean that's the whole point. At 20:30 what will be the jobs? And at the New York conference that I spoke at, it was a great social event after and a person came up to me and said do I feel like Oppenheimer? That is I've invented the nuclear weapon and now I'm regretting that I did it. Well I didn't do it personally. No brand comms invented it.

Speaker A: It's going to happen anyway. So.

Speaker B: Well that, yeah that's my answer as well but I don't know if that's the right answer. I think that it, it's, it is in the wild and it's, it's going like topsy and uh, that's a noxious weed in case international. I think that's the big challenge. Where's it, what are the jobs of the future?

Speaker A: And yeah I'm trying to think as well like um, luckily I'm not a creator in the production side of things even though I do some of my own video editing. But I can see just that there's tools that every week there's a new feature. I'm like oh I don't need that person to do that thing or I can spend less time doing that. And it's becoming just more and more seamless to the point where I'll just push a button, put a prompt in via voice and then it will spit out what I want.

Speaker B: Yeah.

Speaker A: So growth of brand comps AI demand anything you can say, hey, there's a big client that just hey, we can replace this.

Speaker B: Yeah, I can't talk about any clients but I can tell you that the lead time in the US is really an issue for us. But we're only talking to the top 10 advertisers in the world in the US pretty much. Actually there's a couple of variations with our existing clients. Not all of them are in that category but Australia are. Um, the characteristics of the market. The characteristics of the cmo uh, is that they are far less, ah, less risk averse, far um, more inclined to experiment and um, yeah we're having quite a lot of, quite a lot of fun with a uh, lot of CMOs locally.

Speaker A: Why do you think that is? Because maybe the legal risk is a bit more diluted because normally Americans are a bit more um, maybe uh, set in their ways and stuff but also um, I find them a bit more innovative sometimes or more willing to try a new idea even in the risk of failure. I think it's a cultural thing over there but maybe I'M wrong.

Speaker B: Yeah.

Speaker A: In some respects, we're very well established

Speaker B: in the US and Australia, and I wouldn't want to make comment about that, but yeah, there's behavioural differences that are pretty marked.

Speaker A: Business cultural differences, probably.

Speaker B: My word.

Speaker A: Yeah. So large, large companies in Australia are adapting this.

Speaker B: Yeah. The trajectory is that the market, I think within 12 months will be absolutely overwhelmed with offers and that it won't be the creative agencies, um, that have got the really great offers, but the smart buyer will be working out who's got the best rag, which is the Retrieval Augmented Generation, which is the content store. Who's worked out how to build that, uh, who owns it? Where does it reside? Are there issues of copyright? Are there issues of IP protection?

Speaker A: That's different to a dam, a Digital Asset Management Program. It's more of a. What are you calling this? A rag?

Speaker B: Well, the Retrieval Augmented Generation is when the large language models start processing, they go to the RAG before they go to what they've learned from the Internet. So what you build in the RAG will be your content store.

Speaker A: I see. So that's where you put your own, um, content footage and everything.

Speaker B: And then you've got the proprietary versions of the, the agentic AI that are all operating within your ecosystem.

Speaker A: So it's a close.

Speaker B: They're not drawing on the Internet.

Speaker A: Yeah, but they've been trained. Maybe on other things. But then you're not infringing on copyright because everything you're doing is creating.

Speaker B: You're creating.

Speaker A: So the brands that have a big creative repository would be doing quite well then. Um, with that,

Speaker B: I don't think there's too many organizations in the world at the moment that have worked out what we've worked out. So I don't think we don't. We don't see we've got any competitors at the moment in doing what we're doing.

Speaker A: So, uh, when are you raising like a couple of bill?

Speaker B: We'll get around to it.

Speaker A: Yeah.

Speaker B: Okay.

Speaker A: Thanks, Ken. I really appreciate the chat. I think hopefully that clears up maybe where the industry is heading. Um, maybe some people are frightened, but I think, um, you have to adapt and die. Right. So on the bandwagon.

Speaker B: Yeah. Well, it's happening right before your eyes and. Yeah. You know, do it have a movement?

Speaker A: Oh, yeah, thanks.

Speaker B: Nice.