Almost every client conversation I have now lands on some version of the same question. “Should we be using AI for this?” It is usually asked with a mix of excitement and worry, and both reactions make sense. AI has made brand production faster than at any point I can remember in this industry. It has also made brand output easier to mistake for someone else’s. I used to think of those as two separate problems. I do not anymore. They are the same problem, just looked at from opposite ends.
Faster Was Never the Hard Part
Nobody needs convincing that AI changed the economics of content. Copy that took a team a week now takes an afternoon. A voice guide that used to eat six weeks of discovery can get drafted and tested in days. For agencies and in-house teams alike, that is a real gift, and I am not going to pretend otherwise. Budgets stretch further. Timelines shrink. A three-person team can now deliver the volume that used to need a department.
But speed was never the hardest part of branding. Being distinct was. Used carelessly, AI can drain exactly that while making almost everything else easier.
Why So Much of It Starts to Sound the Same
The mechanism is worth understanding, because this is not really about AI being used badly. It is about how the tools work in the first place. A language model is built to predict the most statistically likely next word, sentence, and idea. That is not a bug someone forgot to fix. It is the whole point of the design. So when thousands of brands run similar prompts through similar tools trained on overlapping data, the outputs converge. Ask for “engaging content” or “a bold brand voice,” and the model reaches for whatever pattern shows up most often in its training data, which is also the pattern every other brand asking the same question is about to get back.
None of this produces obviously bad writing. That is actually what worries me about it. It reads clean. It is grammatically fine, brief in the narrowest sense, and completely forgettable five minutes later. I have watched this play out across categories that have nothing to do with each other. A financial advisor’s LinkedIn posts start to resemble a roofing company’s. A wellness coach and a law firm land on the same carousel format. Different audiences, different expertise, same content.
Try Telling Two Brands Apart With the Logos Covered
Try this sometime. Take two competitors in the same category, strip the logos from their last five posts, and hand them to someone who works for neither company. More often than not, they cannot tell you which is which. That used to be a party trick reserved for bad brands, the kind of company that never bothered to figure out who it was. Now it is starting to happen to brands that once had real personalities. The fastest way to write anything is to let a model reach for the most familiar pattern, and without enough brand context, it has little reason to sound like one client rather than another.
I think about it like a signature. Anyone can learn to write neatly. What makes a signature yours is the small, slightly wrong flourish nobody else would think to add. Sand that off in the name of speed and you still have handwriting. You do not have a signature anymore, and nobody recognizes it from across the room.
A large part of brand equity rests on being recognizable without the logo attached and having a voice nobody else could have written. Used carelessly, AI is the quickest tool I have ever seen for erasing exactly that.
AI Can Draft the Brand. It Should Not Decide It.
I am not making the case that anyone should back away from AI. That decision already got made for us, and mostly for good reason, since the productivity gains are real and the tools keep getting better. The question that actually matters is narrower. Are you using AI to produce your brand, or to produce the raw material your brand then works with?
Teams that feed these tools their own data, their documented voice, their real opinions, and treat whatever comes back as a first draft rather than a finished piece, can scale output without losing what makes them recognizable. Teams that skip that step, that type “write me something compelling” and post whatever the model hands back, are the ones dissolving into the noise a little at a time.
The One Question We Do Not Skip
At BrandPipal we treat this less as a debate to settle once and more as a habit to keep. We use AI constantly, for speed, for early drafts, for the production work that used to swallow entire weeks. But nothing goes out the door until it clears one bar. Could this have come from any brand in this category, or only from this one? If the honest answer is “any brand,” we are not finished, no matter how clean it reads.
Here is the strange part nobody prepared us for. AI did not make brands worse at writing. It made writing worthless as a way to prove you exist. Anyone can produce a paragraph now. What used to pass for a voice was often just effort, and effort is no longer scarce. What is scarce is having something to say that a machine could not have guessed. That was always the real asset. We just never had to find out until now.
