Storytelling with artificial intelligence in marketing

According to AI experts, the advancement of AI (artificial intelligence) is still in its infancy, but is progressing at an impressive pace. There’s no doubt that the development of AI is one of the most significant milestones of the digital revolution and, like all groundbreaking developments, it has the potential to make a real difference—for better or worse. Since big things are often reflected in small ones, this duality is also evident in creative content creation using AI in marketing. So-called AI chatbots such as Microsoft’s Bing Chat, OpenAI’s ChatGPT, Google Gemini, and others can easily create seemingly coherent stories based on a few prompts.

The question of whether artificial intelligence can completely take over storytelling in marketing, as well as the advantages and disadvantages of AI in this area, are the core topics of this article.

A brief insight into how AI works

To understand how an AI chatbot like OpenAI’s ChatGPT can create entire storylines in a matter of moments, a brief insight into how artificial intelligence works is necessary.

First of all, AIs are nothing more than computer programs. However, these programs have the ability to evolve by acquiring knowledge on their own. This is done by training the AI systems on huge data sets to recognize patterns using algorithms, and learn from them. This is known as machine learning (ML), which in turn is divided into different models such as decision trees and neural and artificial neural networks. Neural networks are models that use the functioning of the human brain as a blueprint.

Another key element of AI chatbots is natural language processing, which enables the programs to understand and generate human speech. The more a user interacts with an AI, the more the program adapts to the user’s needs. What clearly distinguishes AI from our ability to learn is the incredible speed with which artificial intelligences sift through volumes of data in a matter of seconds, the scope of which is beyond the human imagination.

To visualize: When we research a topic, we spend a few hours reading a few articles, watching YouTube videos, and/or possibly listening to a podcast. In a fraction of that time, AI has combed the entire Internet for information that matches the key points of the topic.

AI's ability to generate stories

A great example of what AI can do for storytelling in marketing is the Afri Cola website (only available in German). The marketing agency behind the site, WALLEK VON ZITZEWITZ Werbeagentur GmbH, had all the content generated by AI, from most of the coding to the images and individual stories of the different Afri worlds. Only the prompt engineering, i.e. the creation of the input commands, and the overall creative concept were created by humans. Users can view all the prompts on the site to understand the development history of each subpage, image and story (Afri-world). The result is impressive. And thanks to the radically shortened time span in which new AI-generated stories are created, the marketing agency publishes new Afri-Worlds on a weekly basis.

While the Afri website can certainly be seen as a positive harbinger of this rapid development of AI in marketing, it’s already difficult to imagine creative development processes without AI. Blogs, translations, graphics, videos and much more are generated virtually at the touch of a button. This not only saves time, but also money that would otherwise be invested in hiring external content creators.

Thanks to an AI’s ability to adapt to the user’s wishes and needs with each prompt, the results it delivers get better and better. The trick is in the phrasing of the prompts.

Prompt engineering for AI storytelling: a brief insight

With a prompt like “Generate a good marketing story for product XYZ that contains 2,500 words”, AI chatbots may deliver a result, but most likely not one that can be used in any meaningful way. As a general rule, clear, keyword-rich, and targeted prompts lead to better results. A prompt should always include the following key points:

• Target audience
• Writing style and tonality
• Purpose of the story
• Form (number of words and/or characters)

Basic prompt tips for storytelling in marketing

1. Define a clear goal

Before formulating the prompt, the goals of the marketing story should be clearly defined. What message must be conveyed? What needs should the story address and what emotions should it evoke?

2. Identify the target group

Every successful marketing campaign is based on sound audience analysis. Strategic storytelling is no different. The first step is to establish who the readers and/or listeners are. This insight determines not only the tonality and language, but also the setting in which the story takes place and the channels through which the finished story is presented.

3. Adapt the conversation style 

Many AI chatbots offer the ability to adjust their conversational style based on the topic at hand. With Microsoft Bing Chat, the three levels are “Creative”, “Balanced”, and “Accurate”. If the content is more factual, Accurate would be the right choice. For marketing stories that appeal to a younger audience, the conversation style should be set to Creative.

4. Use creative prompts

The more creative the prompts, the more creative stories the AI chatbots will uncover.

5. Provide context

AI is very good at finding inspirational approaches based on context. For example, links to external sources such as previous marketing campaigns, company and/or product stories, etc. help.

6. Continuously test and optimize prompts

If a prompt doesn’t produce a satisfactory result, it’s useful to create different prompts and then compare the results. Prompt engineering is a trial-and-error process, especially at the beginning.

Chain prompts for better marketing stories

In addition, it’s important to gradually guide the AI toward the desired outcome, especially for larger stories. A good approach is to create a storyline by asking the AI questions. For example, first ask the AI to describe a scene in which the product to be marketed is shown:

Prompt: 

• Tell me a story about a young man barbecuing in his backyard while drinking a cold beer
• Target audience: 25 to 45 years old
• Text style: funny, relaxed and personal
• Purpose: to promote the ZISCHT beer brand
• Two paragraphs of about 250 words each

The generated story can then be refined prompt by prompt.

Prompt:

• Expand the story: Max tries ZISCHT beer for the first time, bought by his wife, and is unexpectedly impressed by the new brand

With each new prompt, the AI learns to adapt to the user’s ideas and generate increasingly accurate stories. Chain prompting can also be used for character creation, image description, and many other components of marketing storytelling.

Tip: If the prompts are in English, today’s AIs perform better, especially on complex topics. This is because the AIs are primarily trained on English-language datasets. However, it’s likely that the input language won’t play a significant role in the near future.

If you’re looking for a completely new approach to a chain prompt, OpenAI recommends entering the following prompt in the search window and then using the questions to approach the marketing story step by step.

“I would like you to help me create the best possible ChatGPT prompt to achieve my goals. You should ask me about the topic of my prompt and then ask me for more details to help you refine the prompt for GPT-3.5, GPT-4, or ChatGPT.”

The limits of AI in marketing storytelling

#1: AI does not evaluate

Despite the obvious benefits and impressive storytelling capabilities of AI, a healthy dose of caution is needed when using the results. No AI-generated content should see the light of day without careful post-production. There are many reasons for this, but they all boil down to the fact that artificial intelligence lacks human empathy.

ChatGPT, MS Bing Chat, and all other AIs access the data provided to them and derive the results desired by the user. But they do this without any evaluation, neither factually nor content-wise. It’s therefore quite possible that the stories produced contain racist—or at least politically incorrect—elements. The fact is that the AI data sources were all originally created by humans. That mountain of data contains content that has nothing to do with common sense or good taste.

#2: AI isn't a fact checker

Another point to keep in mind when using AI-generated content is that AI isn’t a fact checker. It’s simply a tool that produces content. When it comes to fictional storytelling in marketing, this may not be too important, as in the case of the Afri Cola website. But when it comes to strengthening corporate identity internally and externally, marketing technical products, communicating the brand globally, or launching a start-up, factual misinformation is unacceptable.

These factors lead to the most important tip for working with AI: any content created by AI must be expertly edited. Even though AI increasingly simplifies our daily work, HI (Human Intelligence) is still the key to success.

#3: AI provides similar results for all users

Artificial intelligence has access to massive data sources from which it produces the desired results. AI chatbots such as ChatGPT Plus, Bing Chat, and many others scour the Internet in addition to the training data provided. At first glance, this may not seem like a significant problem. After all, those data sets are huge and growing. But in the long run, it creates a kind of vicious cycle. Marketing content for blogs, websites, newsletters, etc. is being generated by AI at a rapidly increasing rate. Meanwhile, the amount of new content created by humans is declining. This leads to a creative drying up of data sources, which is then reflected in increasingly similar content. This problem is mitigated in the context of marketing storytelling, which tends to involve more complex, creative textual content. However, when it comes to the creation of short content such as slogans, social media feeds or metadata, the AI-generated results of competing market players are relatively close.

Bottom line on AI generated marketing stories

As CEO of Folienwerke GmbH in Winterthur, I am very optimistic about the rapid development of AI in marketing. The potential is enormous and the capabilities of AI are already almost unimaginable. In addition, the use of AI can save a lot of time and money. As a digital service provider, if you close your mind to this, you’ll sooner or later fall behind or even lose out completely. That’s why at Folienwerke we take a proactive approach to AI and train our designers, storytellers, and project managers in image generation, idea ping-pong, and text prompting.

Folienwerke is the leading consulting and design agency for presentations in Switzerland. Storytelling is one of our core disciplines, which we either incorporate directly into the design of slide presentations and explanatory videos, or convey at eye level in our storytelling workshops. In this context, the topic of AI in strategic storytelling is becoming increasingly important.

For us, AI is an ingenious tool that we use for research, among other things. The time savings compared to conventional Internet research via Google & co. is immense. We also use AI to process old blog posts and other text content, to create images as part of slide design, and even to get inspiration for storytelling.

However, I would like to emphasize that HI (human intelligence) always plays first, second, and third fiddle in storytelling. Without human empathy and our sense of humor—in short, without heart and mind—I don’t think it’s possible to create a compelling marketing story that reaches the target audience on an emotional level. And this is ultimately the enormous power that makes storytelling—whether in text form, as an explanatory video, in classic narrative form in a podcast, or as a series of images on presentation slides—one of the most effective marketing tools there is.

Ellen Riesterer is the CEO of Folienwerke in Winterthur. With the leading agency for storytelling and presentations in Switzerland, Ellen has one goal: to enable communication between people in a given time frame with maximum results. Folienwerke offers individual consulting, implementation and training in the field of presentations. With a focus on storytelling, design and rhetoric, Folienwerke guides and encourages speakers to deliver authentic presentations.