How to thrive in the creative AI era : tips from LinkedIn’s head of content solutions, APAC

How difficult will it be for you as a marketer to compete for attention over the next five to 10 years? A lot of marketing comes down to really two big factors, right? Reach and attention. University of Omaha shows that we are overexposed to around 100 ads a day, but about 90% of all ads really fail to do that number one job, which is to get noticed.

Welcome to the Marketing Intelligence Show powered by Supermetrics. the leading marketing intelligence platform helping agencies and brands turn data into business growth. Hi everyone, my name is Dan Halkuli. I lead the content solutions team for the APEC region at LinkedIn.

And I'm pretty excited to be here today. I'm mostly going to be talking to you about how we can thrive in this era of creative AI. And I'm going to share with you some ways that the AI is going to affect our roles and creatives. Hopefully, it will motivate you to embrace AI in both how you approach strategy and also how you promote some of your creative endeavors.

I talk quite fast, so I apologize, but I've only got 20 minutes and this is a massive topic that's moving in many different parts. So I wanted to quickly dive in and let's get started. Let's start with the status quo, where we are right now, which is in the midst of this crisis in creativity. When we talk about this idea of this crisis, it really boils down to a simple question.

So how difficult will it be for you as a marketer to compete for attention over the next five to 10 years? Now, if you've been a marketer for any longer than sort of five to 10 years, I'm sure you would have heard Power Brands Grow, Professor Byron Sharp, Les Burnett, Peter Field, all these guys. And the research that they've done has really taught us that a lot of marketing comes down to really two big factors, right? Reach and attention.

And if you think about attention, attention is really the first step to everything that happens within marketing. Without attention, there's no positive marketing effects for your brand. There's no engagement. There's no recall.

There's no leads. But many marketers, both in B2B and B2C, are suffering from a bit of a tension problem. And we think this is about to get a whole lot worse because of generative AI and their becoming of AI. Now, many of you have probably experimented with AI already.

It's probably making you a lot faster and more productive in your roles. But that speed is also going to mean that there's going to be this massive flood of new content, AI-driven content, coming into the marketplace that's going to be competing against your creative for attention. In the US alone, AI is actually expected to increase the volumes of creatives that we see in the marketplace by about 13% every year for the next 10 years. And that's a tidal wave of content coming to it.

And if I'm honest with you, most of it's likely to be ignored. So research from the University of Omaha shows that we are overexposed to around 100 ads a day, but we can only manage to recall or remember only a few of them, like less than 10% of them. So if we think about that in the context of attention, about 90% of all ads really fail to do that number one job, which is to get noticed. And what determines quality of attention?

It is the creative. It's the quality of that creative that we actually put in there. And unfortunately, creativity has also been fairly undervalued when it comes to how we approach a lot of social media or advertising and marketing the last few years. Talking specifically in B2B, we worked with System 1 and we did an analysis with System 1.

We found that 71% of B2B ads scored just one star or less on their creative score. Now, if you think about that, that means that really three out of all four B2B ads do almost nothing for the business in terms of business effects. And this is because we're all kind of doing the same thing. We're all our creative ways is all of kind of the same, right?

And because it's so the same, it's also a little bit boring. Now, when we say that everybody's kind of doing the same, it doesn't mean that we're doing bad content. the content itself is actually quite good. Everything from fooling up the real estate, punchy headlines, using data in that kind of respect.

And I think that doesn't mean that there's bad content out there. It just means that the quality of quality content has, the bar has been risen. And actually 82% of our decision makers actually agree that current digital advertising is largely forgettable. And that means that this is really huge opportunity for us to stand out and to be interesting.

Now we talk about this as a crisis of creativity, because it's kind of here and now, and it's going to be exploding even further over the next few years with the rise of AI content. So how do you compete? How do you evolve your teams to be more competitive in this new AI landscape? I think this quote from Origin of Species, Charles Darwin, is pretty apt there, that it's not the strongest of the species that survives, but it's those that are the most adaptable.

In order for us to answer this question about how we can best adapt, let's look at sort of the wider context about where this creative AI is really going to play its biggest roles in your jobs. So it really boils down to how we approach the AI revolution. And it really fits into two key areas. One, how do you leverage AI to manage and increase efficiencies in your marketing?

And two, then how do you leverage AI to develop quality creative? Now, if we put that in a two by two, the true north of a future marketing team, future agency or a content team is really going to be able to how you can create and scale content quickly without losing any of its quality in the creative. And if you're a strategist, the same two by two can become about how to use AI to produce faster and more accurate strategies and insights that are also scalable to assist high quality production and targeting.

Now, some of this might seem quite daunting, but if you remain focused on these two core factors, productivity and quality, then it's possible for you to develop a very progressive next-gen creative team. Let's dive a little bit deeper into that at the moment. So let's start with how AI can be leverage across speed and productivity, and specifically around how we think about strategy. So the role of strategists, creative strategists, marketing strategists, it's going to change quite significantly with the introduction of AI.

It's going to pervade everything from your market research to your insights, to how you position the brand, how you target, go to market, how you think about your product strategy. And it's important to think of AI as this tool that is going to do a lot of that heavy strategic lifting for you. Most of the time, we're starting to see this coming out in in the form of AI agents And these AI agents they changing the game in a number of different areas Here are just some that we can pull out for that I seen actually in the marketplace already So the first one category research This is where we seeing people use agents to research the category and then to understand the players, and then sometimes to simplify the category itself, break down specific industry jargon to make those concepts really simple and easy to understand in language that an audience at a mass, sophisticated mass marketing approach is going to appeal to.

On the segmentation side, we've seen agents understand those category entry points and then helping those strategists then build approaches on how to reach most buyers across the most common buying situations. When it comes to content audits, it's a simple now of being able to screenshot a piece of creative, plug it into the agent, and the agent will be able to tell you optimization best practices based on what it's been trained on. Channel and budget allocation, we've seen agents look at real-time making changes to your media spending and targeting based on historical channel surge of periods and current trends that it's currently seeing or going through.

On the competitor analysis side, I've seen an agent understand the race for market share and how it can position real-time mapping of the current TAM across metrics such as share of voice or share of market metrics. And then on content personalization, translation, localization, we talk about a lot of this in creative. This is really about how can we at scale tailor content to individual users using AI tools to make less work for us. And this is just the tip of the iceberg.

There are agents out there that are being built specific for channel management, SEO, for shelf space. I mean, today I just saw an announcement of a company in the US that's launched an entire AI native operating system for the advertising marketing team. On that one platform, they're promising that you can manage your channel strategy, your technology stack, your strategy work. You can create RFPs within it.

You can run your content production in there and also your media distribution. So when we think about this, the future of marketing is really going to be about our roles as managing these many task orientated agents. So in one area that we're already seeing it massively disrupted is in the industry of market research. So in the past, Nielsen and the like, they would have really helped marketers run consumer and market research by interviewing customers and decision makers on qualitative surveys and focus groups and these kinds of things.

This qualitative research is pretty important because it tends to be that guiding light in a core marketing strategy. And it really does influence everything from the four Ps. However, finding enough C-suite executives to complete a survey for a reliable sample has also been found to be really quite difficult, expensive, and very time-consuming when you're trying to do this research. But if you switch this up and you use AI to build synthetic versions of your target audience, then it becomes a viable alternative.

The AI executive does a couple of things that the human executive doesn't do. One, it doesn't get tired. Two, it can answer an unlimited number of questions that you might have about the market and their buying cycles and how they work. And three, it's instant.

It can really be, you can get all this information almost instantly. You don't have to wait for people who attend focus groups and things. And the last part of this, which is really interesting, is that the evidence so far is that a lot of these synthetic AI responses actually have similar degrees of variability to actual human responses. In fact, there was one study that was done by EY, and they conducted a double blind test.

What AEY did was that they provided a synthetic research company who was building these synthetic AI audiences with its annual brand survey questionnaire. And it detailed out its target audience. They were targeting, I think, U.S.

companies around $1 billion in revenue, I think. But what they did was AEY withheld the actual qualitative survey results from the human responses. And then what happened was the research company then created a thousand synthetic personas that matched that target audience's profile and then had them answer the same questions that the human executives were answering. And what they found was that the results were around 95% the same between the synthetic audience and the responses from real humans.

So when we think about the game changing that's happening here around market research, the qualitative market research now can be cheap. It can be instantaneous and fairly reliable, enough to build a strategy from at least. And in the past, we used to have to trade some of these off against each other. Another area that we've seen AI disrupt is on segmentation and category entry points, what we know as CEPs.

I've used this myself where I built a CEPs for a very niche B2B category. It was for IT decision makers who were looking to migrate from SAP's ECC to SAP's S4 panel. If you don't know what any of that means, don't worry, neither did I. But what I did was I basically trained Copilot on the CEPs and what they were, as well as the attributes of the target audience and their problems.

And then I gave Copilot this prompt that you see in the black to produce those category entry points for that category. And then I ran a workshop with the client where I showed these category entry points to the client and we then debated and prioritized which of these positions was the winning play for their marketing team to move forward on. The agents are also in marketing and across other different areas as well. Just here are some of the list of them, right?

Content audits, competitor analysis, creative mock-ups. A great example is the content audits. So in the past, a content audit would often take us a number of hours. Now, after creating a trained agent on LinkedIn's content best practices, I can make a screenshot of any LinkedIn piece of creative.

I can paste it into the agent and the agent will then give me an analysis on how to optimize that creative. As you can see here in the example, I can take it a step further and I can then take that insight and I can engage with a second agent where I can do creative mockups. And here, the GPT here is now offering to do a mock-up for me out of a Canvas API tools, where LinkedIn actually has thousands of templates that can be used on our platform. So I said yes, and then it rendered a post.

I then took it another step further, I plugged it into Sora. And here, I can now make the post animated. And so within a space of about half an hour, I was able to have my creative reviewed, optimized, mocked up, and then animated. Now, it's not perfect, but it is a game changer when you're looking for efficiencies and using AI to make your content much more efficient and your content creation much more efficient.

Another important area where the efficiencies are happening is how to scale certain friction points across content. The big one really sometimes being translation and personalization. So here obviously we can use ChatGPT for the copy and MidJourney for the imagery to help us create localized versions of a single piece of creative And one thing I caution is that while the translations are very useful and helpful they can often fail to recognize things like slang words and cultural terms So it still does require some human editorial oversight when it comes to producing some of these scaled pieces of content.

One other game changing area is that generally I can now scale video. I'll give you a quick story here. I manage the APEC region and I was in need for a content expert in Japan to join my team. If many of you have worked in Japan, you understand that it's quite a unique market and one of the areas of friction is language and translation.

I personally don't speak any Japanese at all, but I still needed to find the right talent in order to join my team and I needed to be where they were, which was to approach them in Japanese. And so I made this. So I'll leave it there. But what I'll note is this is not me.

This is an AI avatar. I made it in HeyGen, 11 labs, and then some post-production in Canva. Now think of the potential scale for this if you are an influencer or a content creator. I can now create dozens of videos a day in any language, on any topic, and I don't need to do any of the time-consuming production stuff, whether it's setting up a camera, doing my makeup, writing a script, all these different things.

Think about this in the context of executive thought leadership. You have an executive who's very, very busy. You maybe only have time for him to speak for an hour and you can get him to create an avatar and push this out. Some of you might be thinking this is quite dystopian.

There's a lot of AI and privacy issues with this. And you're certainly right. These things have not been addressed in approaching it this way. But when we think about efficiency using AI in our teams moving forward, this is a solution that you could potentially explore.

So looking at some of those examples I showed you there today, it is true that if you double down on just using AI tools to make your marketing organization and your outputs more efficient, one thing is quite clear. Some of the quality of what you do is going to be affected. And in fact, we talked about the crisis of creativity and how some creative is very uninspiring right now. This is why we need to balance that speed and efficiency with quality and effectiveness.

And here too, AI can be a bit of a disruptor as well. So these image and video rendering models, as bad as they are right now, that's as bad as they're ever going to be. And in fact, they're evolving so rapidly that they're sometimes taking huge quantum leaps in the space of six months between models or so. And these models are getting better and better.

And they're also becoming a lot more difficult for us to spot in action in the wild. In fact, it's likely that you've already been consuming a lot of AI-generated content, thinking it's human content and not even knowing it's AI content. Two-thirds of the creative industry is already said to have been exposed to AI-generated content at least once a week. And it's increasingly becoming more and more indistinguishable from what we call normal content or human content.

Indeed, only 35% of people are very confident that they can tell an AI ad from a real ad. And as the AI improves, that number is going to continue to shrink and shrink and shrink. And I'll test you out on here. Just in the chat, I would love you to tell me which one of these pieces of creative is not AI.

I'm seeing bottom left. I'm seeing number three. I didn't put numbers against them, guys, but I can kind of guess which ones you are. Bottom left, bottom left.

It's a trick question. These are all AI. They've all been built in Google's latest VO3 model. Now, think about this, right?

This is as bad as the AI would be. And I think that bottom left that everyone was pointing to about the car show, just look at the backgrounds. Just look at the people moving and look how seamless it looks. This is as good as the AI is getting much, much better and faster to the point that if you had glanced at this, you would not know.

So the future of quality content that we will be producing will not be considered about whether or not this was created by human quality or AI quality. It's going to be about whether or not it's effective. We are going to continue to see AI influencing the content creation process over and over over the next few years. Now, we see some brands playing around with this in a tongue-in-cheek way, right?

This push towards AI, this adoption AI is about using it in maybe to make it look purposely bad. And this is a great example from FedEx here where they deliberately used AI because the point they're trying to make in this ad is that entrepreneurs are too busy looking after their own business rather than dropping on the latest AI content creation trends. And so this is one way that you can use it to say, cool, this is the basic way to use AI. There are other companies such as Volkswagen in Brazil that commissioned this campaign where they reunited two of Brazil's most treasured musicians.

It was a mother and a daughter and the mother had passed and they were using through AI to allow them to play a duet together. Now, it was a great ad, but it sparked a very impassioned debate over the ethics of artificial intelligence and its impact to not just society as a whole, but how it's going to transform things like the music industry when you can create AI-generated music. And as we move into this era of AI, we start to see more and more creatives being generated by these AI models.

I do think it's important to note that whatever these models are being trained on is likely also to be what will come out from the other side. A case in point, I asked ChatGPT to tell me how it would identify a piece of content that was targeted for the automotive category. And it said that these were the elements, right? So car hero shots, driving along a beach or an alpine pass, driving in the city, shots about celebrities smiling or a driver smiling and car specific language and jargon and tropes.

This is what the model has been trained to identify is good quality creative in this category. And one might say that this could be seen by the AI as content best practices that this AI model should be using to implement when creating a car ad or an automotive video. But when we think about what is the most memorable automotive ads created, almost none of them adhere to these best practices. And here's the element of caution around using AI on creativity.

Unless we train these models to think more progressively about what works in marketing, the future of content scarily will look like AI written templates of very biased category tropes. So imagine if I had done this and I'd asked it to tell me how to create a good B2B ad, as an example. And then it been trained on years and years and years of very bottom funnel lead generation terrible creative that very rational And I think the reality is we will likely start to see that the AI will continue to produce and contribute to this terrible crisis of creativity that we having with its volume of content that going to come out It going to look very much like that So we're going to drown in a sea of sameness, and it's all largely going to be driven by AI.

This is where we still need humans. This is where we still need to have that. Even though much of this AI can get us on par with human creatives in terms of how the quality comes in, that quality has risen to a part where you are still no longer capturing attention. Remember, we come back to this point of attention.

It's the most important thing you've got to do in order to have effective campaigns. And in order to capture that attention, you need to think outside the box, not within the box of your category. And that's where we seem to recognize that sometimes human oversight and control is there. If we look at the Oxford Dictionary's definition of creativity, it is the use of imagination and original ideas to create something.

So human creativity, by definition, is to push the boundaries to break the mold of what is clearly something that AI cannot quite do right now. This is still the moment where human interactivity needs to be involved. We can definitely get to the best practices. We can definitely have a standard level of good quality, but we need humans to actually push the boundaries, push that AI outside the box.

Some additional areas where you don't really want to surrender your marketing to AI on the creative side is you don't really want the AI to control your brand, its assets, its products, its distinctive brand codes, as well as you don't really want it to control how the brand's people and spokespeople are deployed and used in the creative as well. You still want to have humans oversight on that too. So here's a simplified version of how humans and AI can collaborate across that creative process, everything from pre-production and strategy to post-production and targeting and measurement.

Now, be mindful that there's still legal and ethical considerations when using AI. These tools are super sexy and they're enticing to use, but there are still many different IP privacy data protection issues that you should be wary of. And there's also a lot of biases and hallucinations that we've just seen that these AIs can create based on what they've been trained on. That's where this quality layer must work in tandem with the speed and productivity layer.

AI complements and augments human creativity rather than looking to replace it entirely. And you would leverage that AI for certain aspects of that creative process. And that would be likely to help you improve efficiencies, innovation, and content gets you a competitive advantage maybe, but that human judgment, that oversight still remains essential to ensuring things like quality and ethics and authenticity in your creative. I'm going to close up now with just talking a little bit about LinkedIn's own AI, LinkedIn Accelerate.

So we're a Microsoft company, and we put generative AI into a lot of our technology, and you'll start to see a lot more of it. The idea will be that it will flow through the hands to our members and to our customers in order to help them become more productive and successful in whatever they're trying to do on our platform. And one of the most powerful ways to think about AI, at least how we're doing that LinkedIn, is to think of it as your copilot. It's incredibly capable assistant that's constantly at your side to help you excel in any marketing tasks that you might have, whether that is diving into strategy or creative or doing targeting or measurement.

And with LinkedIn Accelerate, it is one of these tools that you will start to use a little bit more when you jump into LinkedIn's campaign manager. It's an incredibly robust tool and it's designed to do a lot of different things through a number of different agents. And in as little as sort of five minutes, the LinkedIn Accelerate tool will be able to recommend to you an end-to-end campaign and give you automatic optimizations in order for you to reach that right audience with engaging creatives.

Now, how it does is it really brings together your data, like your customer ABM lists and things, or your conversions, and then it combines it with our platform data to help you find audiences and people that are most likely to take an action from the ad that you serve on LinkedIn. We call these predictive audiences. They're using LinkedIn's AI to create this engaged audience that's tailored to your business based on your first party data, some third party data, and some interactions that we see across the LinkedIn platform.

On the creative production side, we do have an AI generated copy suggestions tool in Campaign Manager that's really been fueled by OpenAI. And here it's about leveraging the data from things like your LinkedIn company page to your campaign manager settings, what objectives you choose, the target audience you're after, and then what it will do is it will suggest headlines and copy for you then to jumpstart your campaigns. And as you work through Accelerate, it will be like this co-pilot for you as you look to build more out on LinkedIn support.

And you can now ask it questions directly, such as things like, why did you recommend this budget for me to spend? Or what are the best practices that you recommend I do for targeting this audience? And that co-pilot will then help give you insights in order for you to prove your campaigns with you in control. We're rolling out new features all the time.

So there's going to be some more features coming out in the next few months that's designed to help more on the creative process. It's going to help you with content creation, thinking about creativity, as well as delivering some of those efficiencies when you're developing your campaigns. So if you haven't tried Accelerate yet, jump into LinkedIn's Campaign Manager. There is an option for you to choose it.

And then try out some of the agents that are happening already in that tool and that system. This is just a summary of some of the key takeaways I've talked about. We have this crisis of creativity at the moment. It's going to get bigger with AI coming down.

We want to obviously make sure that everything is, that you're prepared for it. Think about the two by two. Think about how you can use AI to increase your productivity across your team, as well as how to use AI to maintain quality through human oversight. And then have a go at the LinkedIn campaign manager and have fun with it as well.

But that's it for me, guys. It's been wonderful to spend the last sort of 20 minutes or so with you. I see there's a lot of chat and comments in there. I haven't had a chance to have a deep look at them yet, but hopefully some of this was valuable to you.

And thank you so much for your time. Thank you, Daniel. Well, for better or worse, AI is here to stay, but it's always encouraging to learn from others on how AI can actually transform a marketer's role. And instead of fearing it, we should learn to embrace it.

I think we should embrace it, actually. Yeah, look, during inspiration from Daniel, I think I've decided that we're going to replace ourselves with AI avatars for the remainder of this session and pop out for a coffee. What do you think? Nice try, Miles.

but I think the chat is already on to us. So maybe let's move on to the next session.