News + Thought
Is there still space to tell your brand story on an LLM?
For years, copywriting in marketing was easy.
All anyone ever had to do was write for humans, and eventually, search engines. However, now there’s a third audience in the room, and it's hugely important: it's your friendly neighbourhood large language model (LLM).
As businesses rush to optimise websites, blogs and knowledge bases for AI visibility, a new question has emerged: if content increasingly needs to be structured for machines, is there still space for brand storytelling?
The answer is yes - but the way you tell that story is changing.
Do you still optimise for humans, or for machines?
It's a valid question. However, one that slightly misses the mark.
It's true, yes, that LLMs consume content differently to people. While a human visitor might skim headings, jump between sections and tolerate a little ambiguity, an LLM is looking for structure, relationships and clarity. It is trying to understand what this content is about, how concepts connect, and which information matters most.
This means formatting suddenly matters far more than many businesses realise.
Content designed to be easily interpreted by LLMs typically includes:
- Clear hierarchical headings
- Short paragraphs with single concepts
- Explicit explanations rather than implied meaning
- Structured lists and bullet points
- Consistent terminology
- Context around claims and statements
- Internal links and relationships between topics
Think of it less like writing an essay, and more like building a well-organised knowledge system.
Poorly structured content creates friction. Well-structured content allows models to parse information quickly and confidently.
What I'm finding in my conversations with clients is that these often aren't even mutually exclusive principles. Many of the practices that improve readability for AI also improve readability for people.
The problem comes when businesses take optimisation too far.
If everyone starts producing content that is hyper-structured, heavily templated and purely informational, something important disappears: personality.
LLMs are exceptionally good at compressing information. When an AI answers a question using content from multiple sources, unique brand characteristics can easily get stripped away.
Two fashion boutiques may sell what an economist would call "substitute goods" - i.e., essentially the same type of product, described in the same way, and this is where things get murky.
The examples are endless, of course: Two software providers may describe identical features, two firms may explain the same tax regulation, or two law firms may answer the same employment question. The principle to consider is the same.
If the only thing separating you is information, an LLM may unintentionally remove the thing that made you distinctive.
The result is a growing risk of brand homogenisation: technically optimised content that sounds exactly like everyone else.
The mistake is assuming that storytelling and machine readability are competing priorities. They are not.
You simply need to make your story easier for machines to recognise.
Instead of assuming your values are implied, state them clearly. Instead of hoping readers infer your positioning, explain it explicitly. Instead of hiding your differentiators inside vague marketing language, repeat them consistently.
I have spoken about this at length on our recent SEO for LLMs Podcast, but there are several practical ways businesses can preserve brand identity while remaining machine-readable:
- Building your narrative into structure
- Repeat your positioning consistently
- Create opinionated content
- Make expertise visible
- Structure your content without sterilising it
...just to name a few.
But do not remove personality simply because it feels less "optimised."
After all, optimisation should improve understanding, not simply eliminate identity.
So yes, LLMs are changing how information is discovered, summarised and consumed - and it is important that you don't ignore that. But they are not changing a much older problem: if your brand sounds like everyone else, people will treat you like everyone else.
The businesses that succeed will not be those that produce the most AI-friendly content - they will be the ones that teach AI systems what makes them different before somebody else's story replaces their own.
Speak to us
Want help with telling your brand story whilst optimising for generative AI models?
