Like many (many!), I have played with Chat GPT to see how if it lives up to the AI promise and test if it can contribute to better marketing strategies. In this article I share my experience with Chat GPT when it comes to formulating brand ideas.
Marketing about sharing ideas with people to convince them to an eventual (re)purchase. So getting these ideas right is key, especially in a jungle of communication. But before we embark in a communication and media strategy, we need to be clear about what it is we want to talk about, to define what we want people to think about our brand, relative to their needs and the offering of competitive brands.
Chat GPT and Brand Positioning formulations
Having carried out positioning projects for over 60 brands since 2014 and worked with nearly as many brands in my corporate years), I can tell you that “wordsmithing”, i.e. choosing the right words to describe a positioning idea is an important part, especially when you want it be easy to understand for anyone (a.k.a. the granma test) and in any language.
In the first iteration for a famous brand, the system spat an idea that related to the quality of the product being great… This was disappointing because brand positioning is about an emotional benefit being supported by a functional benefit (the quality of a product). So Chat GPT made the mistake many novice marketeers do. Therefore, it needed to be informed with a better frame – which takes us back to the previous article where the quality of the question and the information provided are key.
On the following iterations, I explained to Chat GPT that a brand positioning statement is built around a “brand name” + “a strong emotional benefit” + “aimed a target audience” + “supported by a product reason to believe”. The results improved but remained a bit bland. So in the next iterations I fed the machine with all sorts of brand information, from wikipedia to brand sites and blogs. And to torture a bit more the machine, I asked to provide 10 different positioning statements.
As the iterations ran, I started seeing interesting formulations, i.e. things that were clearly formulated and in line with a brand personality and consumer usage and attitudes. In fact, there was not a single and earth shattering formulation, but several routes that could be worth checking.
Ask Chat GPT to create brand names
I then asked the system to create brand names. If you’ve created brands, you’ll know that finding names that are distinctive, memorable and free to register is a daunting task. Armed with the knowledge that the better the prompt, the better the results, I fed Chat GPT with some brand information (product, history, personality, variants etc..) and asked for names with a certain number of letters, and possibly IP free (although I don’t think the system is able to scan trademark databases thoroughly).
The result produced a lot of garbage – arguably like in a brainstorm session – but funnily enough, it proposed formulations I had come across in previous exercises with no Chat GPT ! In the midst of these names, there were new suggestions that could be interesting. I was on the right path and pushed a bit more the machine torture to provide brand names and variant names that would reflect the brand idea, describe the variant, with rhymes of alliterations. Again, garbage but also interesting results.
Ask Chat GPT to improve copy
As a non-English native (I am French), I sometimes struggle to build simple and grammatically correct formulations. In the (copy) writing field, grammar correction tools are always handy and recent software, like Grammarly, help formulate sentences with more clarity.
And since we already asked Chat GPT to “write a poem about cars using key words from Hamlet”, why not feed the machine with a draft text and ask to improve it for more clarity and impact? Which I did and the result seemed pleasing to read. Or to summarize a text. or to merge 2 sentences… If I were Grammarly, I’d actually worry because Chat GPT can do their job for zero moneys.
In the end, can you work with a smarter partner ?
This experiment cmpelled me to write that this article was not reviewed or polished by Chat GPT. Here, my style, my mistakes ! Errare Umanum Est… Which brings us to the next point : if Chat GPT is good at certain tasks, it would be silly not to use it, would it ? After all, we want to deliver the best.
But if Chat GPT is used, should it not be credited somehow? After all, when we watch a movie, the end credits list all the people who contributed to that movie. But who reads end credits ? Or who cares if a an image was designed with Adobe Illustrator? Nobody. So I suspect that soon we will not care if a piece of work was done with Chat GPT or any AI.
However, AI feels like a threat because it produces great results in an apparent effortless way, using the thing we like to thinks sets us apart (and above?) vs other non-human beings : INTELLIGENCE (*). And in creative disciplines, the idea that an emotionless system can be used to generate emotions (marketing, brands!), can be perplexing.
So, I bet everything you want that in the very near future, you will have “anti AI” movements and we see creative work with a “AI Free” tags, as a way to boast authentic human creativity, just like we also have “Gluten Free” food (or water). Because humans love “authentic” stuff (despite wallowing in fake news).
After this peep into Chat GPT and Formulation, we will have a look in the next article about Chat GPT and Decisions, i.e. provide cues on how to use Chat GPT and the importance on being aware of WHO decides in an AI environment.