Try a simple experiment. Open LinkedIn and click through ten random company profiles. I bet you’ll see at least six shades of navy blue or deep green, three abstract, “dynamic” logos, and one gradient that was supposed to be ultra-modern but actually just strains the eyes.
There’s a lot of irony in this. No one commissions branding just to blend into the crowd. Yet year after year, industry after industry, hundreds of brands flood the market with identical logos and high hopes of conquering the world.
The problem has existed for a long time, and the availability of generative AI effectively exacerbates it. AI models are masters at providing the statistically most “middle-of-the-road” answer to a given question. And in this case, that answer will be a pretty and technically correct graphic – yet devoid of even a single feature that would set it apart from the competition.
“B” as in branding
Let’s start with the basics and explain what “branding” actually is. If you type this word into a search engine, among the top results you’ll find blogs by marketing agencies offering advice on how to combine colors, choose fonts, create a logo, and design a business card.
But that’s not branding! At least not in the sense that has long-term significance for a company.
Branding, understood more deeply as brand identity, is a kind of system. A set of complementary elements that binds the organization’s entire image into a cohesive whole. It includes the logo, colors, and typeface, but also the brand’s character, the way it communicates with its audience, and the archetype behind it. It is not a collection of pretty graphic files to download, but a visual definition of what the brand is (or at least what it would like to be).
The challenge lies in keeping this system consistent. It must stem from something. It must grow from a foundation that was established earlier – namely, the brand’s identity, which in turn arises from a conscious answer to the question: who are we, and for whom are we here?
There’s no accounting for taste – but there is for identity
To highlight what I mean, I’ll use a slightly surreal but real-life example.
I once received a request to change a company’s logo color to match… the CEO’s wife’s kitchen. A family business, a CEO with sentimental attachments – to a certain extent, I can understand that. However, the color in a logo shouldn’t stem from the fact that someone likes the cabinetry. It must stem from who the brand is speaking to, how it wants to be perceived, and what its character is.
This is the root of many misunderstandings between clients and graphic designers. People approach branding from the wrong angle. The business owner comes in with a preconceived idea of a finished logo – they have a gradient in mind, a catalog of favorite colors, and a general “artistic vision.” After a week of work, it turns out that this gradient has no basis in what the brand actually wants to represent. What seemed obvious to the client remains disconnected from their target audience, from the nature of the product being offered, and from reality in general.
We have to start the process all over again. We undo what we’ve done, take a step back, and erase what the client has already firmly planted in their mind. This costs time and money that’s been wasted simply because we started off on the wrong foot.
Branding should stem from research, a brief, and a discovery process involving both the client and the professional. It certainly cannot stem from the CEO’s intuition. Or, even better, his neighbor’s opinion.
Your neighbor is not your target audience
“Joe, listen, I need your help. Tell me, honestly, how do you like my logo?” – and he pulls out his phone, showing a JPG file. The neighbor looks at it for three seconds and passes judgment. It sounds funny, but for many people, this is still the most common way to test branding ideas. The client then takes this to the designer.
I’m not questioning Mr. Joe’s taste. The point is that a neighbor usually isn’t a representative of the company’s target audience. He wasn’t even close to it.
If you’re designing a line of beard care products, your target audience is men who grow beards and take great care of them. Not your clean-shaven friend, not his wife, not their neighbors. It sounds harsh, but that’s how it works: the opinion of someone who isn’t – and probably never will be – your customer has zero value for your branding. (Unless, for some reason, you’re intentionally creating a focus group out of random people and know exactly what you’re measuring.)
„The thing is, a neighbor usually isn’t a representative of the company’s target audience”.
In any case, defining a target audience surprisingly often boils down to “everyone between the ages of 20 and 80.” Of course, that’s not a target audience at all. It’s everyone—which means no one in particular. The more precisely you define who you’re speaking to, the better it is for you and your branding. Colors, typefaces, the form of the logo – these are all decisions a designer makes based on what they know about the brand and its audience. Without this foundation, we’re left with nothing but the subjective aesthetic preferences of individual people.
I remember a client who came in with elaborate visual plans and, essentially, a comprehensive vision. He even had a finished app and had purchased elements directly related to the new branding. On paper, it looked like a project that was at least halfway there, but when we started talking about the fundamentals – who the product is actually aimed at and how it wants to be perceived – he didn’t know what to say. I call this the zero point. We supposedly already have a logo design and an app, but the brand still can’t present itself.
Don't just fix the symptoms. Start with a diagnosis
A Logo in 30 Seconds
Of course, artificial intelligence comes to the rescue today. It’s hardly surprising. Working with a designer and the entire discovery process involve real costs, especially for a brand just getting started that barely managed to fit the product into its budget. We all have ChatGPT in our browsers. We type in “make me a logo that fits a pallet-hauling transport company” – and in 30 seconds, we get something. Practically for free, fast, no conversations, no process, and above all, no tedious questions from a nagging graphic designer.
I understand the temptation. The trouble starts when the client (who isn’t a designer and doesn’t see the difference between a logo and a brand) gets some result from the AI and lacks the expertise to judge whether what they received is good. Will the logo work at different sizes? Will it look good in a monochrome version? Does it actually fit the industry? Will it resonate with the target audience? The client doesn’t ask themselves these questions because they usually don’t even know they should be asking them.
You can’t file a complaint with ChatGPT. If the logo doesn’t work and you’ll regret it a year from now – you have no one to turn to. A human designer takes responsibility for their work. They’re someone you can go back to, discuss things with, who justifies their decisions and explains why a particular solution – whether it involves color or graphics – doesn’t work in the context of your brand. We’re not talking about personal taste here, but about arguments based on experience and knowledge. AI models work quickly and correctly, but they lack that “human” background.
For my part, I must also add that every graphic generated by AI carries a certain visual baggage. I guarantee that the trained eye of someone knowledgeable in the field can spot it. I don’t know when this will fade away or at what point the line will cease to be noticeable, but for now, it’s there.
This doesn’t mean, of course, that AI is completely useless. I know of clients who have effectively used these models, even if just to seek inspiration. They didn’t know beforehand how to describe their vision precisely, so they typed snippets of ideas into the chat and saw what came out. It wasn’t a finished design, but they got an interesting starting point for a conversation. That’s a sensible way to use the new technology. It’s worse when such a starting point is mistaken for the end result.
It’s beautiful… statistically
Personally, I hardly ever use ChatGPT in my work. I don’t generate references, I don’t ask for concepts, and I certainly don’t throw in random prompts hoping for a flash of inspiration. Everything I design starts with a sketch. I build a logo from start to finish, let’s say “the old-fashioned way.” It may be less efficient and take a lot of time, but it’s my process, and it yields something an algorithm can’t provide: an idea. My own, concrete, creative idea, based on what I know about the client and their audience.
If I do use a chatbot, it’s for completely different purposes, such as searching for information, organizing data, or sometimes generating simple texts. For creative work and graphic research, definitely not. I do my research manually because that’s where you might accidentally stumble upon something that doesn’t quite fit, but contains a single inspiring element.
The model won’t spit that out because it treats topics too strictly. Meanwhile, truly creative compositions are rarely literal.
In other words, there’s a difference between generating and inventing. It’s worth keeping this in mind, especially if we aren’t graphic designers or creators. Regardless of the quality of the prompt, the AI model returns something statistically probable that fits the pattern. Good branding, on the other hand, is precisely about transcending that pattern; ensuring the brand doesn’t look like every other one in the industry. This requires a person with experience and a dash of ingenuity, allowing for non-obvious combinations.
Interviewing the Client
Now that we know what branding is (and what it isn’t), let’s move on to the practical side. The creative process begins with a brief, and we start drafting the brief with an honest conversation about what the brand is meant to represent.
A brief shouldn’t be just a form to fill out during a coffee break. A one-on-one interview works better. Sometimes it’s long and quite difficult, because clients come with ready-made answers that turn out not to address the most important questions. I ask about the target audience; they say “everyone.” I ask about the brand’s character – they describe the product. I ask about the archetype – they look at me as if I’d just switched from speaking Polish to Romanian.
„A brief shouldn’t be just a form to fill out during a coffee break”.
It’s not easy, but it’s precisely from these conversations that the things emerge which later guide the entire project. It’s in the brief that we discover the brand has a completely different audience than the one the CEO imagined. That a product which looks like something from the premium segment is sold mainly in the mass market. Or that “young, dynamic brand” is a self-description that says absolutely nothing.
Why is this so important? Take Peugeot as an example. Its logo features a lion. Not an engine, not a wheel, not a road – nothing to do with the automotive industry at all! The lion is meant to express a certain attitude. To be associated with strength, agility, nobility, and perhaps prestige. A coffee brand with a coffee bean logo simply describes what it sells; yet it could tell a story about what it is and what it stands for.
The Genetics of Branding
This is what the entire process of discovering identity is all about. It’s not about creating a “pretty logo,” but a logo that embodies character. That’s why I only select colors, fonts, and form based on what I learn about the brand. It’s a sequence that must not be reversed or shortened. Meanwhile, AI essentially skips it entirely. A chatbot won’t ask you what you really want to achieve. It will generate something that statistically matches your description.
AI answers the question “what?”, while a designer answers “why?”.
For this very reason, I always encourage decision-makers to participate in the process of creating the brief and conducting discovery. People who know the company and its DNA best. If the brand doesn’t know its target audience, market research must be organized. Surveys directed at actual customers, external analyses – these are tools that are important for sales and marketing departments, but they also provide the designer with a foundation for solid work.
However, no designer will do this work for the company. Their task is merely to ask for the results, analyze them, and translate them into a recognizable, effective brand.
Good branding doesn’t start with a logo, but with a conversation. If you feel your brand is ready for this conversation – we look forward to hearing from you.
