Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode 001 - Kellyn Coetzee
26 Plays4 days ago
Welcome to the very first episode of Mind the Model: The Modern Marketer's Guide to AI! Your hosts, Emmalee Crellin and Nathan Guerra, kick off their journey into the AI rabbit hole by sitting down with none other than Kellyn Coetzee, Head of Digital at Zenith Australia and a leading voice in applied AI.
This episode is packed with practical tips, ethical considerations, and a visionary look at the future of marketing with AI.
Get ready to revolutionize your approach to AI as Kellyn shares her expert insights on:
• Navigating the "AI Black Hole": Discover how Kellyn tackles tricky AI challenges, bouncing between tools like Anthropic artifacts and Gemini to build mini-websites and overcome token limits.
• Mastering the Art of Prompting: Learn Kellyn's two-pronged approach to AI prompting, including the powerful RISEN framework (Role, Instruction, Steps, End Goal/Context, Narrowing) for structured outputs, and even her secret "bribery" technique to get the best results!
• Democratizing AI Skills: Hear Kellyn's passion for making AI accessible to everyone, shifting roles towards more strategic, collaborative, and client-facing work.
• AI's Impact on Creativity: Unpack the role of AI in advertising creativity, from scaling AB testing to optimizing variations, and why the human element remains irreplaceable.
REFERENCES:
Character.ai: AI-powered platform that allows users to create and interact with customized digital characters through text and voice. Users can chat with characters based on fictional figures, historical personalities, or even popular culture icons, as well as create their own chatbots
RISEN - the format that Kellyn uses for prompting
More information can be found here: https://clickup.com/general-resources/playbooks/ai-prompts
- R - Role: Tell the AI who it should be or what expert persona it should adopt.
- I - Input: Give the AI the specific information, data, or clear instructions it needs to work with.
- S - Steps: Outline the exact sequence of actions you want the AI to perform to complete the task.
- E - End goal: Clearly describe the final result or the specific format you expect for the output.
- N - Narrowing: Add specific constraints, rules, or limitations to make the AI's response more focused.
News that was mentioned in the chat:
https://www.pbs.org/newshour/politics/vance-rails-against-excessive-regulation-at-paris-ai-summit
AI For Everyone - AI Course that Em mentioned:
https://www.coursera.org/learn/ai-for-everyone
And for the record, Claire Polkinghorne is Australia's most capped footballer, with 169 appearances.
Transcript
Introduction to Mind the Model Podcast
00:00:09
Speaker
Hello and welcome to our first episode of Mind the Model, the Modern Marketer's Guide to AI. I'm Emily Crowley and I'm joined here today with my co-host Nathan Guerra. So once a fortnight, everybody can look forward to hearing from the two of us as well as a special guest.
00:00:23
Speaker
We'll be talking to CMOs, agency leaders, consultants, and probably some a influencers, eye influencers, AI influencers. I think I was just a little sick in my mouth saying that. Just workshop that one. Yep. Yeah, yeah.
00:00:34
Speaker
but Let's just say we'll be talking to smart people throughout the industry who understand how they're using ai and I think where they go to learn a little bit more stuff
Hosts' Passion for AI and Marketing
00:00:43
Speaker
as well. Yeah, definitely. And just a big context and setting the scene for our listeners in our first episode.
00:00:49
Speaker
How did we get here? Why does the world need another podcast? but You and i have known each other for about a year now. And over the past six months, we've been going out for so many coffee meetings, really just diving into the AI rabbit hole, meaning to touch some grass after our conversations because we're just going so deep into it.
00:01:05
Speaker
and But we're not the experts. We're just really insanely curious about what AI is and the powerful machines that power it. and We both are so passionate about the world of marketing and advertising.
00:01:16
Speaker
and that intersection with ah AI, Gen AI as well. So we thought the podcast would just be a great opportunity to kind of have these conversations and then also shut up and listen from really smart people. So yeah.
00:01:27
Speaker
Yeah, absolutely. So the idea is each show, we're going to be doing this learning stuff. We're going to be playing around with some tools as well along the way. and I guess seeing it's taken us about six months to pull all of this together, I thought it would be probably good to the level set.
00:01:41
Speaker
Understand and share with the world, what have you been working on with AI in the last six months, Em?
Emily's Experience with ChatGPT
00:01:46
Speaker
Over the past six months, I've gone really deep into some tools. Specifically for me, it's ChatGPT, but it's really brought about so many efficiencies in my life, both personal and professional.
00:01:56
Speaker
From the professional side of things, I look to it as a really handy writing co-pilot. It's kind of helped me in my journey of reaching out to new people in the network, kind of helping that imposter syndrome, helping grow my confidence to yeah find the best way to articulate what I want to get across. And so I find that writing...
00:02:13
Speaker
assistant quite helpful, but I mean, in my personal life, I also use it from the perspective of helping plan for a recent trip away or also fixing my running toilet by uploading pictures to chat GPT asking what's going wrong with my toilet.
00:02:27
Speaker
Yeah. Yeah. It's funny i because we used to use, and just go straight to YouTube for that, but now you've got an AI assistant to help you along the way first, right? That's exactly. And like in my most recent example of fixing the toilet, it did send me YouTube videos that I was able to watch, but it's just that one-stop shop that gives you everything.
00:02:43
Speaker
and But what about you? What have you been up to in the past six months?
Nathan on AI in Business Scaling
00:02:46
Speaker
Well, having recently started my own business, I'm finding that I'm wearing literally every hat possible, which is fabulously fun and and and exciting and joyful.
00:02:55
Speaker
But it does mean that I have to constantly do everything. So I'm using it to write copy. I'm using it to write Python scripts. I recently scraped a website. I'm not to name exactly which one.
00:03:08
Speaker
And got 4,000 leads using a combination of a Python tool called Beautiful Soup and some code. I don't know how to use BeautifulSoup, nor do I know how to code, but luckily I do know how to use gen AI. So it's been really helpful for me in terms of managing and wearing all these different hats. And I think that's one of the things that I'm potentially most excited about with AI is its ability to help small businesses scale incredibly quickly.
00:03:34
Speaker
You've seen businesses now with worth a $100 million dollar average run rate and 10 people Yeah, love that. And for you, do you find that you are asking your friends as much or are you ah really leveraging just being that startup mentality, entrepreneurial mindset of kind of going first to Gemini? Well, look, i yeah, I do love Gemini. I guess I still have those Google roots deep in me, that bias.
00:04:01
Speaker
But yeah, I do find that i can I can do a lot more than I did before without having to go to friends. I used to go to friends a lot more. I guess now I'm more of a
Introduction to Kellyn Coatsy
00:04:08
Speaker
a loner. Maybe I don't need friends anymore. Living the life of the recluse.
00:04:12
Speaker
i I think it's time we shut up and listen to who our first guest is but Why don't you go ahead and introduce the clever cookie that is Kellen Coatsy. I would love to. So I'm going to do, I've got a little bio for you, Kellen here. Are you ready? So Kellen Cote is the head of digital at Zenith Australia.
00:04:30
Speaker
ah She's co-chair of the IAB AI Council. She's a mentor at the Global AI Institute. And at Publicist, she spends most of her time driving performance marketing innovation by translating emerging technology into real world results.
00:04:44
Speaker
And with over 15 years of experience, she's a leading voice in applied AI, which I'd love to know more about, and really known for helping teams unlock scalable brand-safe solutions. So I previously have known Kellyn and we worked together quite closely when she was head of ai and insights at Kineso.
00:05:01
Speaker
And she really does blend for me that kind of ability to communicate incredibly complex, clear ideas or ideas and make them clear and engaging. So I can say that I have known her to be a real standout leader, speaker and mentor.
00:05:14
Speaker
And I'd love to welcome our very first guest, Kellyn Coatsy. Hello. Thank you guys for having me. I love em how you said, I think you like you ended up on recluse and like, And that speaking of recluse, let's just bring in Kellyn. You are who are I go to when I think of recluse now.
00:05:31
Speaker
probably no I love it. but No, thank you so much for that. warm Welcome.
Kellyn's AI Model Preferences
00:05:36
Speaker
Helen, just to kick things off as a little bit of an icebreaker, what is your preferred gen AI model right now and why?
00:05:43
Speaker
So thinking chat GPT, Gemini, Claude, Croc, what's your preference? What's your top one? It's really hard because I use all of them for different reasons, but I would say the one that I use most frequently would be chat GPT for its reasoning models.
00:05:56
Speaker
Okay. But you do jump bounce back and forth? ah Yeah. Just, it really depends on the task, what I'm doing. but like, to be fair to chat GPT, I would say thats that wins in frequency. Okay. Yeah, love that. And is it paid?
00:06:08
Speaker
um No, I've got my own subscriptions to a bunch of different models. I do claim back on tax. Love it. Tip for our listeners. um I think from a work perspective, we have a lot of proprietary ah tech and AI things that we should be using in work context, which I do.
00:06:25
Speaker
So, Kellen, I would love to know when was the last time you fell into the AI black hole?
Overcoming AI Challenges with Kellyn
00:06:31
Speaker
And let me explain that. This is another idea I'm workshopping, like, eye influencers or whatever.
00:06:36
Speaker
The AI black hole is when you find yourself back and forth with an AI for, like, 30, 40 minutes just trying to get to a solution that works. You know, you can you're constantly like, I know it's there, and you're almost close.
00:06:49
Speaker
Yeah. What was the last time? I've been playing with the anthropic artifacts. So, yeah. I was trying to find a shortcut on for work on how to put scopes of work together.
00:07:00
Speaker
So I ah want to build a little mini website for myself where i can just take on a bunch of things like a case of this client once, advanced analytics, propensity modeling, coupled with an SEO, whatever, and then just click certain buttons and it builds our scope of work and I do it from there.
00:07:15
Speaker
the challenge that I got to with Anthropic is that while it was really clever and intuitive and it made an awesome interface, the JavaScript functionality behind it kept on bumping into my token limits.
00:07:27
Speaker
So for those listening, token limits, just imagine like space, you get by like 200,000 tokens per query in a chat. and That's what I did. is I eventually copied the code as it was within Anthropic, popped it into Gemini with the 2.5 with million tokens.
00:07:44
Speaker
yeah um And then I was just like, fix this. And then it did. And yeah, so I think that took me longer than it probably should have. Yeah, it's funny. I had a very similar experience um where I was trying to use a combination of Claude. And I find that the Claude, particularly the free one, has such low limits in terms of what's
Effective AI Prompting Strategies
00:08:01
Speaker
allowed that I just end up back in Gemini.
00:08:03
Speaker
Yeah. Because the free version of Gemini still has that huge context window. Yeah, 100%. So it's really powerful. Yeah, I was using it. I think you bring up a good, sorry, just to jump in. I think you bring up a really good point though, Kellen, of sometimes the tool that you're using is not ideal for the task. And so sometimes bouncing around to different tools who will actually get you there in the end. Yeah. I think it also comes to that point where I think when things become a lot more agentic, oh, sorry, not agentic, when we hit AGI, when one particular model will have multiple different functions, that will be a lot easier. But for now, I have to use a combination of different solutions to get to specific outcome that I want, where if I'm
00:08:40
Speaker
putting together a presentation for a client and I've got a slide that I need to use an image for. And the image is very abstract, like, i don't know, like laziness. I've got to try and instead of like going on there, instead trying to find an image, it's just easier to generate that image, but I'm not necessarily a creative person myself.
00:08:58
Speaker
So what I've done is I've used ChatGPT's reasoning models and I've built little GPT for myself where I just put in a concept. It gives me three conceptualized ideas of that concept in terms of like an art an artistic idea, I'll pick that one. And then it translates that into a perfect image prompt for Sora, which includes like resolution, turn all those different things. So it's the combination of different things to get to that outcome.
00:09:24
Speaker
That's a perfect segue into my next question, which is going to be around like, how do you approach prompting? Because... I find that often, and and and you know this isn't just me talking about this, but using an AI to help you write the prompt that you need for
Understanding RISEN Framework
00:09:37
Speaker
the right tool is is one way. But I'd love to know a little bit more about what it is and how you get the best out of prompting AI.
00:09:45
Speaker
I teach that there's two different ways to prompting. There's prompt chaining, where you go back and forth with AI. It's conversational, it's creative. You don't quite know exactly what the outcome you want is because you're asking questions along the way.
00:09:57
Speaker
It's great for strategy, for creativity. thank But if you need a more structured output, like for instance, when I'm teaching people in our agencies to write insights from from the reports, the insights usually need to be something significant based on observations that lead to ah an actionable recommendation optimization. Like that is what it's That's a formal structure. 100%.
00:10:20
Speaker
hundred percent So for that, i will I use a particular framework. there There are many different frameworks out there. I use the most comprehensive one because why not? Because you're just going to do it once, save it, and then just use that GPT over again.
00:10:34
Speaker
The framework that I use is Risen. So give it a role. which i know, Nate, you have opinions about. Give it a roll. Give it an instruction. Tell step by step exactly what you want it done.
00:10:45
Speaker
And then end goal, which is essentially context. And this is what I love about AI. So actually studied psychology and linguistics and I think, you know, language models and the way that language, how AI learns from language. that There's so much human behavior in language and the things that it's learned from books and where things have been written.
00:11:04
Speaker
If you've ever in your career been given task but not been given the context of why you were doing that task, often you overcook that that task or you undercook it. You never quite nail it then.
00:11:16
Speaker
AI is the same because it's learned from human behavior. and when you give it the context, it will give you a much better result. So like I think that's just so fascinating, that parallel between human behavior and you know this technology. So role, instruction, steps, end goal, which is context, and then narrowing. That's where you can say, be concise, try your best.
00:11:35
Speaker
Here's $1,000 good result. i I do find sometimes bribery does get the best out of it in a weird way. It's learned from human behavior. So it's like, if I often like if i really want a good result, if I'm being very pedantic, I'll say, this is very important to me. Please try your best.
00:11:51
Speaker
And I'll often get a better result. Some great tips there. So risen and bribery is what I took away from that. Yes, correct. ah Excellent.
00:12:03
Speaker
So Kellen, you and I go way back. um I think we pretty, I think we first met in San Francisco. yeah When i last worked with you, you were, I believe your title was like queen of AI at IPG, something like that.
00:12:14
Speaker
Oh, I'm so sure. But how did you go from like head of performance or kind of that role to an AI role?
00:12:25
Speaker
Like what was the process like? Yeah, a
Kellyn's Career Shift to AI
00:12:28
Speaker
great question. I think it was a bit of luck and a lot of preparation. Okay. So for those of you in the in the industry will know, if you are an activation specialist, especially working within the Google stack, a bit of meta- You are very close to the Google brain, the google deep mind technology that is powering AI. If you have bought search campaigns between, well, me since 2007 to now.
00:12:55
Speaker
So i think that prepared me, the fact that I studied linguistics and psychology and that's embedded language models, I think there was just a natural interest. I'm a bit of a gamer, you know, building gaming PCs.
00:13:07
Speaker
i I know a lot about NVIDIA chipsets because want to play Dota without lag. Like it's just, there's so much right nerd and tech stuff there. But then I had the very fortunate situational opportunity where at the time with IPG, Kineso, Metakine, Reprise, Merge, and there was a lot of restructure and a lot of opportunity.
00:13:26
Speaker
ah that we spoke about like different roles, GM, et cetera. yeah and what But what would I actually want to do? And I kind of put this role together like a bit of a, if I was ordering a pizza and I had amazing leaders who who believed that I could do it, backed me, saw the importance of it and had that vision to push it through Jess White, Andy Holford.
00:13:44
Speaker
They really advocated for it. And then it became a role. And then like Anna Theorese also massive and cheerleader for that space. Now that your AI is embedded into you, what you do, how are you using it like broadly and your current
Role of an Applied AI Specialist
00:13:59
Speaker
role?
00:14:00
Speaker
ah I would describe myself as an applied ai specialist. So how do you apply AI to any role? So whether I'm ah using the tools to make certain tasks faster or better, there's that application of AI.
00:14:14
Speaker
Then there is a profound understanding of how all of the partners like Google and Meta, and Trade Destin, how are they using ai and what does AI mean to them and their systems? What can it do, can't it do?
00:14:25
Speaker
I think as a marketing specialist, we all need to understand that really, really well. But I think... I'm sorry, that learning of the partners, is that on you to understand the partners or is the onus on the partners to be able to upskill their agencies?
00:14:39
Speaker
When the upskill comes through, it can often come through more of its like a sales pitch, like at all the wonderful things that you know AI can do. What are you talking about? I never did that. Never. Never.
00:14:50
Speaker
from where I started back with search, we had like four things we could target keywords, location, device, and like time of day. Now there's more than 200 different signals at any given moment, including like lifetime value chain. Like there's just so much there and you can't dispute that.
00:15:05
Speaker
But what I can and will dispute, we're using generative AI to create creatives, but I can't input my clients guardrails and restrictions around that.
00:15:16
Speaker
um That's a little bit too dangerous for me to play with. So while The platforms might be incentivized to push those particular products. I think it's up to an agency to understand what those limitations are, understand what the AI is capable of.
00:15:29
Speaker
I would say it's beneficial to understand how AI works. Like for me, if I understand how something works, like like it that makes everything so much easier. And then tying back to your time at IPGA where you said you went on a bit of a two-year crash course, were there any specific courses you took? What kind of education or was it just kind of playing around with the tools and understanding it more?
00:15:47
Speaker
There are some amazing courses that if you do like a Coursera, also did like a Harvard Graduate Computer Science 101, which was and not AI specific, but I think that that grounding really helped.
00:15:59
Speaker
I follow a podcast, a guy who used to own a marketing agency, and then he turns to do AI education and applied AI education. So a lot of his examples and theory is in my field. So i actually got so much from them and they are the artificial intelligence Institute.
00:16:18
Speaker
The links are hu all in the show notes. like the gens Yeah. I was just thinking that. Yeah. The guys that do that, they're brilliant. And then lot of little LinkedIn courses of here and there.
00:16:29
Speaker
I think I really struggle once you get to a certain point, it's very difficult to learn something new beyond the basics. So that's when I started turning to a lot of the AI leaders in the field, started following what they were talking about, what that what excited them, what they were arguing about, and then just using AI to figure that out or like explain this to me, why is this a big deal?
00:16:49
Speaker
um And then Yeah, ah eventually. So it's like learning about AI and then using AI to learn. Yeah, I love that. And I know you're really you're quite passionate about wanting to share that knowledge with others and to help educate people, really democratize the skillset of AI.
Educating Colleagues about AI
00:17:03
Speaker
So how do you approach bringing other colleagues on that journey? And from your perspective, how much of it is just training versus building up processes? Yeah, I love it. Like, When I do a workshop and I can see people are often very nervous or fearful. And then we start talking about it and learning about it. And they're like, oh, it's actually quite accessible. Like I can do this.
00:17:21
Speaker
It's not scary. Yeah, it's not scary. And like, I love that people think I'm really smart. I'm not actually like much smarter than than they can do. They can do all the things that I've learned. And like when they, read I can see that light click. And then when they get it, like,
00:17:33
Speaker
This can actually change how I work for good. That is like one of the best feelings to see that that moment. So I would say a lot of the the the training that I do is understanding how IAL works, how they can use it safely within the tools that we Within the guardrails of the agency, yep. Yeah, and like what tools do we have in our company to allow and facilitate that.
00:17:53
Speaker
For the tools that we need outside of that, how to use them without... using sensitive information or opening ourselves up to risk. Um, cause think it's a combination of both, like until, you know, you can get everything you need in one space.
00:18:07
Speaker
It helps to know how to use different tools. So yeah, so what it is, how to use it then in terms of process, I think there's, I've got like a few universal AI applications that people can use for everything, but I'm trying to go really broad, whether you're in HR, finance, activation strategy, how does everyone use things?
00:18:28
Speaker
But i'm more specifically now with my role at Zenith, as the head of digital, I want to have an AI excellence in our process to digital. So how does it integrate into our reports? What technology are we using to help Speed up the process. really weaving it into the DNA of the culture. Exactly. But for me, that is like, that's for me, an added bonus. The passion comes in. It's just, I want people to understand this new language because once they can speak it, they can also build these things themselves. Like they don't need me or the CEO had to come in and get them. It's quite empowering. Yeah.
00:19:03
Speaker
Yeah. yes um And then from your training, have you noticed a generational difference in adoption or interest? Yeah. oh Was that meant to be pointed? It's funny.
00:19:15
Speaker
Well, the the reason I ask is because I was at a recent ACAM um event where they, think it was someone from The CMO of Iconic stated that ironically, a lot of the older Gen X of the company were really bought into
Generational Attitudes towards AI
00:19:29
Speaker
it.
00:19:29
Speaker
yeah And Gen were quite apprehensive because they were utilizing Gen AI as cheating. So they were as on board with it as the older colleagues. So yeah, just can't get your point of view on that. We old people are happy to cheat.
00:19:42
Speaker
I think that's so fascinating. i think you mentioned like three things on there that can be like an hour conversation. So this, I think the way that educational system or like universities, schools have first responded to AI as a ah tool for cheating and things has probably instilled to this kind of like fear and shame for using it.
00:20:04
Speaker
But I'm starting to see now how AI is actually empowering educators and, and you know, people are learning differently. So I think that will change. It's fascinating. And I can see why some of the um Gen Z coming up might might feel that way.
00:20:18
Speaker
There is, um there is, data and studies to suggest that younger people use ai more for like as a therapist as a social conversational agent whereas millennials will use it more for work whereas boomers i don't know if they use it much at all from my perspective when i do training i found initially maybe two years ago a lot of apprehension from some of the older the generation interest from gen z but like Still, I wouldn't say that they had any skills higher or like they weren't using it more or doing anything more than anyone else.
00:20:54
Speaker
These days, now that I've been doing this for two years, I worry when I go into ah training for a younger group, am I actually going to teach them something that they didn't know already? But luckily every time I do a training or a workshop, people come up afterwards and say, oh, that's so interesting. I didn't know that. And I'm, i I'm genuinely surprised that I'm teaching some relatively basic things that people still don't know yet, which is slightly concerning, but also It means I have a job for at least another six months or so.
00:21:19
Speaker
I want to try and just distill that. do you Have you noticed any difference between the younger people and the older people? It sounds like it's it's pretty similar though. I mean, like they both need training. like what What are the things that are different?
00:21:31
Speaker
So both need training. With the older generations, it's more like, okay, that's very interesting, but I still just need you to do that for me. And that's not what I, one. right They want you to do the work. But also, yeah, but that my whole thing is more like,
00:21:45
Speaker
I'm trying to teach you how to cook. If I just made you a meal each time, that's not one, a good use of my time, but also like anyone can then come in and cook your job. Like it's just, you know, like, maybe we'll check back with you in six months and see if the answer is any different.
00:22:01
Speaker
Oh, the thing i was going to jump in and say, and you brought up a similar point about your time at IPG where the, the senior leadership is really bought into applied AI and how it would benefit. There does seem to be the companies that do quite well in the space of integrating AI into their workflows. It comes from that top down and bottom up approach where the bottom, you know, the the entry level employees really generate the ideas and the use cases for incorporating AI into their work. And the leadership is fully bought in and endorsing of it at a board level.
00:22:32
Speaker
and How would you find in your experience and that meeting in the middle happens? Well, I think I would challenge that hypothesis to begin with, because I think when it comes to process and embedding AI into process, i would say it's a lot more automation than then AI, for one.
00:22:47
Speaker
And I think that's also a lot of the challenge and in my role is people like, how do I use AI to make time sheets easier? Yeah, so sometimes it's more like ah an automation thing, which is like MarTech and a whole other world that I know maybe just the the tip of the iceberg about.
00:23:02
Speaker
So that's one side of it. when i When I talk about the leadership support that I've had is that backing and echoing of how important it is to understand, again, I'm going to use the metaphor, how to cook.
00:23:13
Speaker
Because we can, you know, the the agency can have this this kitchens and all these ingredients and they can spend so much money in setting that up. But if at the end of the day, you guys don't know how to you actually use it apply it to your work, then then that's a bit of a problem. And what I, I've structured my workshops and the things that I do very specifically. So when I teach, this is how AI works, this is how you build a specialized GPT, this is how you do this, this is how, and then I go through this very specific use cases for ah products or craft, personal development, all these different pillars.
Importance of Practical AI Examples
00:23:48
Speaker
Every example that I use, I've specifically chosen because it's not something I came up with. It's something that someone came up with because I told them how to use AI. And then they came up with that idea.
00:23:58
Speaker
I'm like, cool, that's a great idea. I'm going to show other people how to do that. And I think that is really powerful. So you're just basically stealing from other people.
00:24:07
Speaker
Look, ah you've had a ton of experience, Kellen, across multiple holding companies, but multiple countries. I know that you've been in Australia during the Gen.I. i or the AI revolution, if we're going to call it that.
00:24:18
Speaker
But I'm wondering if you're noticing any differences between countries, like what and how Australia kind of compares in terms of its adoption of things. Well, statistically, Australia are massive AI adopters.
AI Adoption in Australia vs. New Zealand
00:24:32
Speaker
What I found really interesting though. So I think our average Australian is using AI between seven and eight minutes a day. Marketers are using it up to 240 minutes a Whoa. whoa Yeah. So we are like really embedded in our own sort hype cycle and like leveraging it. And it makes sense. Like if you think about people, the biggest AI players, how do they make money? It's through marketing. so where are they going to test and play with AI? It's, it's in our industry, it's in our field. So it makes sense.
00:24:58
Speaker
and We get that kind of exposure. I recently had the privilege of going to go down to New Zealand to um do like a bit of an AI talk. I always turn these future of AI talks into actually just very pragmatic, guys, this is how it works, this is how you should use it. Like just use it as like a way to democratize the skill.
00:25:18
Speaker
And all the conversation I had with some like really big brands and really senior people, it felt like they were about 18 months behind where a lot of Australians are. And I just felt like They are so keen and desperate to know this stuff.
00:25:31
Speaker
And I'm so keen and desperate to teach them. What functions within the business are seeing the greatest efficiency gains at this point? Hypothetically, if the business intelligence or um data analyst department were leveraging AI and machine learning, the hours that they would have spent gathering data from different um platforms, putting it one space, cleaning it, visualizing and analyzing it,
00:25:55
Speaker
absolutely AI would help so much with that process and they can spend more time the analyzing, finding insights. And then they start to become more like media strategists and, you know, think
AI's Impact on Business Functions
00:26:03
Speaker
about it from that way. So yes, that is one particular function.
00:26:06
Speaker
But would I say that that AI is then saving the company so much more time at that? Like, I just think that AI shows up differently in so many different departments in different ways that it's really hard to compare. Okay. It's more about like,
00:26:21
Speaker
If whatever role I'm in in the business, what is my business problem? What are the tasks that are taking me too much time? What can be AI? What do I know about AI's capabilities to apply it to that problem?
00:26:32
Speaker
Okay. So everybody can see the benefits of AI. Yeah, 100%. And i like there's a guy, Andrew Ng, he was the founder of Google Brain. He said the same thing. It's like, you will get more value out of ai by giving and democratizing that skill set to people than you would in building AI solutions from scratch.
00:26:50
Speaker
hundred percent agree and i love Andrew Ng because I actually took his Coursera course on generative AI and it was fantastic. and Just going to shift a little bit here in your current role, we've talked a lot about applied AI within the group itself, but
Client Curiosity in AI's Media Impact
00:27:08
Speaker
are there clients coming to you for applied AI upskilling?
00:27:12
Speaker
i think the big question clients have is how is AI impacting media and marketing? And within that answering that question, like, you know, very obvious example is like how AI is impacted user search behavior.
00:27:26
Speaker
People are using more voice search, more visual search. They're searching in GPT, they're searching social. It's completely disrupting the space. And, you know, clients expect us to have an opinion about that. And we're building out solutions to do that and building out those solutions.
00:27:41
Speaker
When I do speak to clients, I find it very but beneficial to take them along that journey and say, okay, this is the AI are we're using. This is how it works. This is how the teams are using AI. like And through that, I think there's a lot of interest like, oh, you guys are doing some pretty cool things. Can you come and share my teams? And i there is an such an appetite for that, which is ideal for me because that's the the the type of AI i love to do and teach is the applied.
00:28:06
Speaker
I love it. And to me, it sounds like it's more so solely focus on like the marketing and the media lens, not so much as you know, their specific business problems that they might be coming to you with.
00:28:17
Speaker
Yeah, column A, column But I think we have the authority to speak about, you know, those marketing pieces. And I think when you you gain the trust, you can show we understand the technology, we understand the impacts, we've got these solutions.
00:28:30
Speaker
Then I think that gives you the credibility to talk to other business problems that and become more consultants and advisors for for brands on the aren and internal. Love that. and I'm going to throw it over to Nate for possibly the spiciest question of today. How worried are you about AI replacing you as a stand-up comedian?
00:28:50
Speaker
o I mean, I never really was a very good stand-up comedian to begin with. so Yeah, pretty well. But you know what? Actually, so my my mom's husband reached out to me and said like, I'm doing a presentation tomorrow. Can you give me a joke about AI? I'm like, dude, use AI. And then, yes, no, my first response was like, no, can't joke about AI. I take it very seriously.
00:29:17
Speaker
But i I think the, when I, and I know this is like a silly joke, like a silly question, but also kind of very sincerely, i think, The art of comedy is also in storytelling, ah taking people on a journey.
00:29:30
Speaker
When I speak to an audience, I'm trying to engage and and see if they're connecting and bring them with. And I depend so much on them for timing. Am I saying the right things and reading the room? It's that human element you bring. Yeah, you can't you can't replace that with but AI.
00:29:46
Speaker
Well, that's great because I'm also interested in understanding how that applies more broadly to creativity in advertising as
AI's Role in Creative Processes
00:29:53
Speaker
well. yeah I've actually had this discussion a few times this week, interestingly enough, i chatting to a lot of people from Leo's and Saatchi and Saatchi. We're very topical here.
00:30:02
Speaker
Yeah, yeah. No, but just ah in hidden the publicist building, like there's just such like a heavy weight of like creative genius surrounding And I think I'm the first to tell you that I know a little bit about AI, but I don't know everything. And I think when I speak to creatives about AI, I'm learning far more than I'm teaching when it comes to that.
00:30:22
Speaker
From what I can tell you as a performance marketer, when it comes to creative, the ability to be able to take a really good creative and change the headline and CTA to do some AB testing is fantastic. And to be able to do that at scale, amazing.
00:30:36
Speaker
Is it going to dream up the next ah Christmas campaign for like a grocery person? like i don't I don't know. I think that- I mean, what did you think of Coke's last Christmas campaign, right? Yeah. Look, if I were to get those performance creatives and change the CTA, I but i like I'll have opinions about like if if I'm getting a creative from a creative agency, if they can make it so that we can use Gen A within platforms to alter certain things. So is that, does that mean then having the products front and center or to to a particular side so we can change things?
00:31:08
Speaker
Like, I think there little elements of that that I can advise, but I mean, my mom has been dressing me for years. I have no taste. I have no creative like talent. So I'll leave the creative up to them. But if I can,
00:31:21
Speaker
if we can work together to find that safe space to optimize, AI can be really powerful in the productization, like, yeah, im just doing it at scale. And what do you see the future agencies to be?
00:31:32
Speaker
Genuine optimism. So, and the reason for this is because I actually built myself a well um specialized GPT. Also, sorry, just a little side quest quickly. There's no universal term for GPTs, gems.
00:31:46
Speaker
Yeah. Artifacts. Yeah. Like it's very. What would you name it? Apes. Apes. Okay. okay Why not? Does it stand for anything? Well, it's like AI apps.
00:31:57
Speaker
Build this ape where either submitted a job description or just even a title and we'll break down that particular role into all the different tasks and say of this task, AI can do maybe 80% of that, 60% of that, 40% of that.
00:32:11
Speaker
And what are the tasks that AI actually can't do well? What are the skills someone needs to excel in those particular things that AI can't do? What tools they need to use for AI to do the things it can do.
00:32:23
Speaker
And what is the evolution of that role in a future where AI can do some of those like more boring tasks. And every time I've put in any sort of description in the marketing universe into this, the future of that role is always more strategic, more collaborative, more client facing. And like all the stuff that we actually want to be doing, we want to be doing timesheets and like ah data cleans and reporting, like all those different things. But we want to be doing the analyzing, the insights, finding, applying insights to creative ideas. And that for me makes me super
Future of Strategic Human Roles with AI
00:32:58
Speaker
excited. And I've seen it firsthand when I've actually gone through that.
00:33:02
Speaker
Yes, but there's a lot of arguments that people have become very good at their jobs by doing a lot of that boring stuff. Em and I were talking earlier about this, and she was saying you know she wasn't able to use AI for a particular task, but she knew the data...
00:33:18
Speaker
inside and out as a result of having to spend the time doing it. Yeah. and Are we not, are we fading a generation of marketers who are growing up, but not understanding how things work yeah because they're able to use AI or they're being skipped entirely? um Yeah. I would turn you back and say how things worked before. Okay.
00:33:39
Speaker
So for instance, in that ah data conversation, i've I've been through something like that myself, where I was trying to build a dashboard for client by using all this data, happened to stumble upon some really interesting cross category insights whilst doing that that tedious process.
00:33:54
Speaker
But if I had spent less time pulling data changing all of the a dollar numbers to just dollar numbers like we're doing all of that stuff could i have spent more time with an ai with that data investigating like hey do you have any insights that's not really an interesting insight to me because that's probably because we went live with tv this week here's the media plan tell me something else and like the interaction that we would do with that data when we had the time to do that would be significantly different and i know it's a bit facetious but i'd say that like the way that we used to build houses with like spades and, you know, like things and how we build houses now.
00:34:32
Speaker
It's the same, right? i mean I mean, was going to say grave digger, but then I thought that was a bit more good, but like and equipment to do different jobs. And as a result, we're able to build like skyscrapers and do like very different things. It's opened up so many different types of jobs because of the advance of that technology of the equipment that we have to do these things.
00:34:51
Speaker
But I think that, While we might be a little bit worried about how we do things today, I don't think we spend enough time dreaming of all the things that we will be able to do that haven't quite just imagined yet. The talent and the employees that will get on board with Applied AI and have that AI-first mindset will rise to the top.
00:35:11
Speaker
similar to how those who are really keen on digital and keen to learn and be a part of that will rise to the top and it will reshape the industry yeah in a beneficial way.
00:35:21
Speaker
I hope. Yeah. And I hear it ad nauseam and I kind of like exactly what you said. I believe it wholeheartedly, but we keep hearing like those who know how to use AI will, you know, not replace, but like AI won't replace you. It's those using AI, et cetera. Yeah.
00:35:36
Speaker
But, And i do I do believe like when I see people who can grasp and utilize AI on how they are shortcutting tasks and doing work and thinking, managers thinking like directors and like it's just, it's so inspiring.
00:35:51
Speaker
But if you're listening to this and well, anyone who hasn't quite been, like it's a journey. If you're at the beginning of the journey, don't be intimidated by all the things you you feel like you don't know. It's actually relatively easy once you get going.
Learning from AI Failures
00:36:04
Speaker
just Just dive in. Don't be scared of it. Don't avoid it. just to wrap up, because I can't believe our time's almost ending with you, but and thinking about a time different from the black hole scenario, but but we like to talk about the AI wins, right? Of, you know, unlocking efficiencies and leveling up managers, thinking like directors. I love that line.
00:36:23
Speaker
But for you, can you tell us about a time when you had an AI fail? It just, it absolutely did not work for you. What did you learn from it? The times when I've been caught, using AI. But I mean, again, I didn't find it a fail. Like, you know, one too many em dashes in that email and you're busted. Like it's just, I put in like spelling mistakes on purpose. So when people talk and think about AI, they're thinking ChatGPT or Gemini or what, maybe can you give us one niche product that you think is really useful that maybe people aren't aware of?
Security and Responsible Use of AI Tools
00:36:58
Speaker
ah think the other the i've yeah i've been getting a lot of joy out of playing with the anthropic clawed artifacts a lot of what i teach um and when i do workshops and things is that even some of the most simple applications of ai comes with a lot of risks if you don't know how ai works or how this application works and there are a lot of bad actors and nefarious opportunities that people can then infiltrate like secure information all these different things so i think yeah As much as I encourage people to go and learn and play, i think we also have responsibility ourselves to to know the dangers and understand the limitations and things. And I think, yeah, we don't talk about that enough.
00:37:40
Speaker
Yeah. Yeah, great. Well, look, I think that's time for our final question and may be an easy one. Do you see yourself as an AI optimist, pessimist, or realist? Ooh, probably optimist.
00:37:52
Speaker
i would like I would like to say a realist. But from what I've seen and what I see through the eyes of some of the the AI leaders that are close to the future of this technology, and I see that their optimism, I think I'm probably a lot more optimistic.
00:38:08
Speaker
Well, Kellen, thank you very much. Thank you so much, Kellen. No, very welcome. This has been absolute pleasure. and yeah, I can't wait to tell you how things are going in six months. Thanks again, Kellen. And for everybody else, remember the intelligence might be artificial, but the wins are all real.
00:38:24
Speaker
See you next time.