Reality-Check: AI-Tools in Design & Marketing

Ready or not?

Let's be honest: Originally, this was supposed to be a text containing tips and tricks for the everyday use of AI tools in design and marketing tasks. All à la: "The first hype about AI is over, let's summarize how AI inspires us in our daily doing." As a preliminary interim conclusion, however, we have to say that AI – despite all the promises of effectiveness and efficiency – has not yet become indispensable in our day-to-day agency work. But why might that be?

Between gimmick and FOMO

While we see ourselves as innovators and early birds in other areas (e.g. the development of complex MACH architectures or the creation of creative customer experiences), we are probably more of an early majority when it comes to AI. Our inboxes fill up every week with newsletters about artificial intelligence; we (occasionally) play with new, free tools, exchange tips on the right prompting in workshops and (some of us) follow AI news via podcasts and industry socials. But while on the one hand it is of course super exciting to observe what is happening in the field at such a rapid pace, this does not mean that we are actively utilising the potential of AI.

By the way, our exploration of AI also remains rather playful because often we have to work with test data and exemplary use cases. After all, not every project has AI potential from the outset; and understandably, not every customer makes their data available for experimental purposes.

Tools in abundance

In our critical self-reflection on why "the artificial intelligence" is not yet our employee of the month, we realise that AI has far too many faces. And sometimes we can't see the wood for the trees. This may be due to the fact that AI tools (or those that pretend to be) are now available in abundance. New AI tools are popping up every day, either covering super niche use cases or claiming to be the solution to all of humanity's problems. It is honestly difficult to separate the wheat from the chaff – especially without having an insight into the technology behind it.

And even if a tool sounds promising, it's not always so easy to try it out. You often only get access via exclusive invitations or only for short trial periods, after which you have to pay directly for use. This actually makes it difficult to gain experience in the first place and discover solutions that really offer added value.

Simplicity vs. complexity

Once we have found AI solutions that we test intensively, we realise repeatedly that blind faith in their results is dangerous. In addition to data protection issues (which would be worth a dedicated article), the tools currently available require precise instructions. Fast results are often just a prompt away - but their quality often still needs to be improved and should be critically examined. While AI is strong in processing large amounts of data, many generative AI solutions that rely on user input only scratch the surface. Simple tasks such as creating agendas, image sets or HTML snippets work perfectly, but AI tools often fail when it comes to more complex requirements such as detailed wireframes or designing new features. How do we explain this?

Interpersonal factors, experience-based values, dependency networks and implicit requirements that do not exist as data points are crucial here – because AI cannot fully capture these aspects. Although AI is perfectly capable of processing complex data, it also requires correspondingly complex inputs. If these are missing, the results are disappointing. Or to put it bluntly: "Shit in – shit out". (It may be that AI will improve in this respect and this passage will age badly :-P).

And yes, in all honesty, perhaps self-satisfaction also plays a role: at least we still like the result of our own work better than the result of some AI. And holding on to this feeling somewhat nostalgically, somewhat egotistically, is also quite nice in times when there is increasing speculation about the future of creative professions.

Tools that we like and use

Despite all the external (and internal) hurdles, there are still tools that we already use for tasks in conception, design and digital marketing – even if some of them are not used on a daily basis or across the board. These are not only solutions that rely on AI, but of course also tools that support us with automation. As a reminder: Generative artificial intelligence uses existing data and processes it to generate new content or collate information; automation in software, on the other hand, helps to carry out repetitive processes automatically. So both are smart approaches, although the first is of course a little smarter.

Here is an incomplete list of tools and extensions that are particularly popular with us in the Creation and Growth team (with the exception of SEO tools that have natively integrated AI or machine learning, such as Termlab).First and foremost, of course - how could it be otherwise – ChatGPT. Most tools are useful and take more and more work off our hands, but we could still do without them.