Marketing and AI: a realistic perspective as of 2024

Adland and (mostly) all other marketing people are going bonkers about AI since the launch of the first chat-GPT Version over one year ago. We see the use of generative AI in visual context that replaces manual work of graphic designers but also in other areas of the marketing field like branding, strategy and many more. Almost all areas of the typical agency workflow is touched be current developments in this area somehow.

With every hype there come dozens of different opinions on future scenarios, where one see the end of the agency model coming and others see the AI hype passing by. The truth, like so often, lies somewhere in the middle: there will be a decent influence on the creative industry but at which timeframe and intensity? A more critical view on the topic has as usual Mark Ritson, to which I can somehow relate as well:

In most cases, the big global brands are experimenting with AI for the same reasons they experimented with the techno porn pinup of 2021 – NFTs. Ageing CMOs don’t want to look out of touch.

https://www.marketingweek.com/ritson-differentiation-brands-ai/

So let’s break down some aspects of the current AI in marketing discussion:

Creativity and craft at risk…

It didn’t take long after the generative AI powered image creators launched and improved to acceptable visual results, that brands and their agencies hop onto the trend and created AI campaigns with creatives partly or fully crafted by AI. Of course the first mover advantage generates a additional buzz in the marketing scene, but at the core the AI campaigns must also deliver to the consumers. Coca Cola’s AI generated from the future soda product failed miserably to deliver actual value to consumer in our present and was heavily criticized. So my current perception is, that we are now beyond of this first wave of curiosity and first mover buzz driven AI infused campaigns and creative work and get back to our basic task: to deliver value for the client through creativity.

AI generated Video can create realistic looking scenery which would cost a tiny percentage when comparing it to traditional shooting processes. For the very cost intense video content production industry AI could really mean a huge change. Already now studios shoot only a fraction of the scene in real places rather then in front of a green screen in a huge studio. Series like The Mandalorian are entirely produced in studio with all backgrounds generated by computers. This is already the practice now, and AI can elevate this even more by cutting costs drastically for the imaginary creation. Also actors in Hollywood are already concerned that they will be replaces by avatars and the actor is only needed once for creating the 3d model of her/him.

Another aspect is the integration of AI tools in the actual craft process like image editing, texting etc. Adobe for example is putting a lot of efforst into the integration of AI to its complete product range. Is that a good thing? – I think yes. Because really time consuming tasks, like for eg. masking off hairs of people in photoshop or put object in front of different backgrounds, are now easy to realize with the AI integration. This saves up time for the more sophisticated creative tasks like ideation etc. One aspect that agencies need to be aware of is that the saved time and money will on the mid and long term again be claimed by the clients in the form of cost cuttings.

Higher, faster….and more productive?

AI will be a productivity boost compared to the era when the personal computer was introduced to the offices in the world – this is what at least the big 5 consultancies claim. To put into context: these companies step into a huge new consulting opportunity and of course they only give the best forecasts for AI implications. Is this realistic? – nobody can tell, as not a single person on this plant can predict the future. Do I think the productivity gain promise is overdrawn? – Absolutely.

If you have used Chat GPT or any other form of generative AI, you will get easy with a simple prompt to a decent result. Is the result good enough or even the correct answer to your questions? If it gets advanced and you really want exact things or continuity in imaginary, it gets quite difficult and text-intense to realize this. Prompt engineering is already claimed a new field as the input for AI can get very complicated. To give context, give expectation the role are just a few aspects that pro prompters use. A prompt can easily have then a few hundred words. Prompting in advanced use case for me therefore is ineffective and I hope that the interface to AI will see some more improvement in the future.

VC Andrew Chen recently published a quite progressive view on AI and marketing in which he stated Infinite labor, content, interactivity as a consequence of the AI usage. He is having the perspective that marketing will be completely personal in the future with million of 1:1 communications and therefore be stronger coupled to sales. This is a very martech driven view on the topic which I can somehow understand, but I personally don’t think this will come to life widespread. Even if we have the technical possibility to create dozens of pieces of content tailored to every individual, the critical point here is still, will the consumer enjoys it and sees a value behind it. This perspective I miss the most in current discussions of AI – the one of the consumer of the AI content.

Why we tend to forget the consumer in the AI hype?

It sound really as the paradise for marketeers: with AI we can easily create high volumes of content at a super low cost that is tailored to my consumer. But: consumers don’t like to consume normal ads – why they should like to consume AI ads?

I think that is the most critical question in the entire discussion about AI in marketing, but is rarely discussed as everybody assumes personalized content will be just more appealing by definition. What then happens to the big brand building campaigns with emotional stories, like to Coke Christmas truck? Does everybody get their own personalized Christmas truck ad then and how will this resonate? Or how will the consumers react if every brand comes up with a version of a Christmas truck? When AI enables the creators of ads to almost infinite output, the consumer will as well quickly create mechanisms themselves to avoid or filter those, because their ability to consume communication will stay the same as now – uninfluenced by AI.

Only the technological possibility to create something, makes it not useful and valuable to the consumer. As Mark Ritson also states in his article, the call for additional differentiation gets louder and louder in a world of overloaded AI ads and content.

We have to just be honest to ourselves: consumer will also not like most of our AI ads, just by nature because it is advertising. That’s reality but we have developed profound mechanisms as our creativity, emotional storytelling entertainment etc. to overcome parts of this effect and create effective campaigns.

Democratizing, really?

I remember back when social media started, I was part of this thinking that it will democratize the way media is produced and distributed. Do I still think that way – definitely not. I was blue eyed and very young back then and later reality took over and gave us social media empires like meta with more in-democratic power then ever, that have proven terrible impact on our society. I think with AI it will be all the same and the statement of democratizing something is just a nice saying – I am not really sure WHAT exactly we possibly democratize with AI: the generation of generic content, access to crawled information of the internet? So for me this statement is just promotional with no real foundation behind it. What will happen is that the power of AI will be in the hand of just a few companies that control the the base models and the super expensive infrastructure to run AI – and this is a scary outlook. Of course this will be the big GAFA – and newly joined by OpenAI and Nvidia.

Welcome to the bubble

Recently Scott Galloway made a point that we are currently quite sure in an AI bubble, that is comparable to the one from the dot-com era. I mean there is some true evidence in it and I think Scott is absolutely right here. But comparable with the afterlife of the dot-com bubble crash, there will be a big market for AI tools, use case, infrastructure etc. on the long run. Should a bubble be a reason to not tap into the field now? – Absolutely not. I think AI is here to stay, whether there is a bubble or not. One tool I always relate to when evaluating a trend or hype is the hype cycle from Gartner.

As you can see Generative AI is at the top of the curve, which means at the peak of its hype and will later settle back to a normalized level over time. So knowing this, gives a clear picture of what will happen in this field and also some indication on how to react and prepare. So, don’t fall for the hype 😉

Links


Comments

Leave a Reply

Your email address will not be published. Required fields are marked *