A professional writer’s view of where AI belongs in the content process, where it absolutely doesn’t, and why knowing the difference matters more than ever.
schedule
Read time:
short_text
Word count:
I have been a writer for over a decade now. Long enough to have worked through several emerging threats to the profession, innumerable changes in digital marketing and more declarations of ‘the death of long-form content’ than I care to remember.
Most came and went without troubling me too much. But, man, AI has been different.
When free AI tools burst onto the scene, it was immediately obvious that here was not simply another quick way to check spelling, suggest synonyms or shave a few minutes off cleaning a transcript. Here was something capable of producing an apparently competent article in seconds. I could give it a brief, and it would mimic a tone and deliver 1,000 reasonably well-organised words before I had finished making a cup of tea.
For those of us who make a living from writing, that might have been the moment to back away slowly, cut our losses and throw in the towel. But right from the start, I went the other way.
Maybe it’s because I forged my first career, as a digital consultant and then an in-house strategist, in the fire of the internet. From my first job at an advertising agency, working in its ‘New Media’ team (four of us in a single hired office in the Truman Brewery on Brick Lane), to my last corporate role leading a massive technical migration project, I have always been more than comfortable embracing technology (Reader, she married it).
So, I started using AI early because I wanted to understand it properly. I wanted to suck up all the benefits. I hated it for trying to steal my job but understood that the change it brought was as inevitable as death itself. And I was not going to go gentle into that good night.
I threw myself into ChatGPT with reckless abandon. I think my first prompt was: ‘Tell me why you are better than me, a freelance human writer.’
So began a fiercely competitive, challenging and ultimately hugely rewarding friendship.
After an emotional start, I became scientific and methodical. I tested what AI writing tools could do well and where they became unreliable. I looked at how they changed their approach when given more context, better instructions or stronger source material. I built custom GPTs and tested them over and over, watching how easily they could drift from the original instructions or repeat mistakes that had gone unchallenged within a conversation. I stopped, refined the instructions and started again.
Once I had control and consistency, I began to use AI on real content projects, using the tools to surface research I could validate, test different structures until I found the strongest route and check spelling. Immediately, they made my process faster and my work broader, more accurate and, ultimately, better.
I began to understand my own value more clearly.
The very useful colleague with no lived experience
AI is extraordinarily good at dealing with volume. So much better than humans. Better even than the 1,000 monkeys typing Shakespeare on their typewriters.
Even simple LLMs can analyse vast amounts of information and pull out themes and key points, compare several documents, sort a chaotic collection of notes and help impose order on a sprawling project. Using dictation, I can walk my dog and talk and talk to ChatGPT as I yomp through the fields. When I get home, I can ask it to take everything I have shared and turn it into emails, reports and even treatments for the pieces I want to write next.
That is no small thing.
So much of professional writing happens before the first sentence is written. Its shape forms in the writer’s mind as we consider the arc of the story, how to unpick the complexity and how to lay out its component parts so the reader can understand them.
Then there are sources to find and validate, research to absorb and assess, background to understand, terminology to check and a mass of potentially useful material to shape. On a complex project, the difficulty can often be holding all of that information in your head while you pick out the thread that will draw it together and make it meaningful to somebody else.
AI can carry a great deal of that cognitive load. It certainly gives me more space to think.
It can take all my disparate thoughts and ideas and hand them back in a nice package.
I use it to interrogate research, explore possible structures for an article and spot connections I may want to investigate. It can turn an unruly brain dump into something I can assess, group source material by theme and show me where the evidence feels thin. It is also extremely useful when one substantial piece of thinking needs to be adapted across several formats.
None of this is about pressing a button and accepting whatever appears.
Working well with AI is an editorial process. Every output has to be questioned, checked, reshaped and, sometimes, wholly rejected. The skill lies in briefing well, maintaining ownership and recognising when you have something genuinely useful versus something that simply looks finished.
Because AI has a remarkable ability to sound as though it knows what it is doing.
Good, modern professional writers have developed the equally useful ability to notice when it does not.
Competent writing is no longer enough
The first wave of AI-generated business content was so easy to recognise. It was full of ‘unlocking potential’, navigating ‘rapidly developing landscapes’ and embarking on ‘transformative journeys’. Everything was fast-paced, ever-evolving and, for reasons nobody could adequately explain, a game-changer.
Of course, the rash of em dashes was the killer giveaway.
The technology, and the apparent ‘intelligence’, has improved significantly since then. Terrifyingly so, maybe. With carefully considered prompting, plenty of context and a strong editorial hand, AI can now produce cleaner and more natural copy. The obvious verbal tics are disappearing.
The deeper problem, however, remains.
AI works by drawing on patterns in the material that already exists. Broadly speaking, it generates language by predicting the tokens most likely to follow what has already been written. This is why so much unedited AI content feels strangely familiar. Even when you have never read that exact sentence before, you have probably encountered every part of it many times.
It is the archetypal competent writer. It is smooth. It is grammatically and stylistically correct. It contains all the expected points in roughly the expected order.
You read it, it makes sense and you may even learn something.
Then you come to the end of the piece and realise that nothing in it surprised you, amused you or delighted the bit of your brain that sits up and pays attention when it falls upon a clever, twisty bit of writing.
“I could hear the clock, and I could hear Caddy standing behind me, and I could hear the roof.”
William Faulkner, The Sound and the Fury
“People are trapped in history, and history is trapped in them.”
James Baldwin, ‘Stranger in the Village’
Neither of these sentences is memorable because it follows the statistically safest path. One of the reasons they stand out, is because they don’t. Faulkner repeats words that an efficient editor might remove and asks us to hear something that cannot literally be heard. Baldwin turns his sentence back on itself, transforming an abstract observation into an unforgettable trap.
A machine could arrange words in any of these ways. What it cannot independently supply is the lived experience, artistic intention and instinct that made these particular choices necessary. These writers knew when unexpected construction would reveal something that conventional language could not.
Good business content does not need to imitate literary fiction, but it does need a reader to hold and engage. Like fiction or speechwriting, it needs rhythm. It needs judgement. It needs to know when a detail will make an abstract argument jump out at the reader and show why it matters. It should be able to vary the pace, find an unexpected turn of phrase and recognise the point at which another paragraph of explanation will make the reader Lose. The. Will. To. Live.
Most importantly, it needs character.
Content should feel like somebody is speaking to you
The best business writing reads like a conversation with someone who knows their subject and respects your time. The writer is present. You can feel intelligence, warmth, curiosity or conviction coming through the words. The turns of phrase, the vocabulary and the pace tell you something about the organisation even before the copy makes any claims about its values.
This human presence matters so much more now that every company, any company, can generate an endless supply of tidy, grammatically correct content.
I hate to burst your cost-saving, CFO-placating bubble, but a distinctive brand voice can’t be created by adding ‘make it conversational’ to a prompt.
Conversation is the playground of personality. This is where your values and truths and approach come out to explore and have fun. It includes emphasis, timing, instinct and the occasional phrase that would definitely fail a review by committee but is exactly the phrase the reader remembers.
Honest human content also reflects choices.
One organisation may explain a complicated subject with reassuring patience. Another may use sarcastic humour to puncture the pomposity surrounding its industry. A founder-led business might sound candid, energetic, and slightly impatient with conventional wisdom. A large corporate might ground its words in years of evidence and experience. These qualities grow from real people, genuine beliefs ,and an understanding of their unique audience. They are part of the brand’s identity.
To some extent, AI can help a writer express identity consistently once it has been properly understood and articulated. It cannot independently decide what makes the organisation interesting.
The value sits in the thinking
There is a persistent idea that using AI somehow reduces the authenticity of the finished work. I understand this concern, particularly when so much automated content is published without any meaningful human involvement.
But using AI carefully allows me to spend more of my time on the parts of writing that deserve that human attention. I can devote more energy to the interview that uncovers the real story behind a project. I can look more closely at the tension within an argument, find a stronger opening and worry over the one word that changes the tone of an entire paragraph. I can concentrate on whether the content is honest, useful and recognisably connected to the organisation publishing it.
I can test and strengthen an article’s structure, then spend my time finding the original story that will bring it to life.
The value of a human-led approach to writing lies in understanding what needs to be said, noticing what is missing and shaping the material so that another person wants to keep reading. It lies in deciding which detail brings a story to life, and which one merely gets in the way. It lies in hearing the phrase an interviewee uses in passing and knowing that it contains the heart of the whole piece.
Those are editorial decisions. They are creative decisions. They are also commercial decisions, because the quality of those choices determines whether the content makes the brand clearer, more trusted, and more memorable.
The AI search paradox
This combination of human thinking and machine assistance is incredibly important in AI search. Generative search systems need enough clear, relevant content to understand what an organisation knows and when that knowledge may be useful. Clear structure, direct explanations, well-labelled sections, and self-contained passages all make information easier for a machine to use. AI can be extremely helpful in creating and checking those elements.
But visibility also depends on having something worth retrieving in the first place.
Thousands of organisations are now using the same tools to produce content from the same publicly available information. Inevitably, much of it converges around the same ideas and language. The same statistically predictable choices. This adds ever more repetitive material to the web. AI search systems may then retrieve and summarise that material, producing answers shaped by the same converging language and ideas.
And so it continues, spiralling into a black hole of beige sameness.
Another generic summary of a well-covered subject gives an AI search engine little that is new or distinctive to retrieve. What creates value is credible source material that contributes evidence, experience, specialist knowledge and a clear perspective. Original research can do that. So can a detailed case study, an expert opinion, a useful comparison or an honest account of what happened when a company tried something and it did not go according to plan.
All of these spring from the fertile bed of human experience. The irony is quite satisfying. AI has made it possible to produce an unprecedented amount of content, while simultaneously increasing the value of the qualities it finds hardest to imitate.
Loving AI without handing it the keys
So, no, I do not worry that AI will make professional writers obsolete.
It will continue to change the work. AI has already removed much of the value from writing that is generic, formulaic and produced without much thought.
I am comfortable with that.
I’ve never seen my role as being to fill an empty space on a content calendar with an agreed number of words. My role, indeed my raison d’être, is to understand complicated things, find the interesting thread and communicate it in a way that sounds clear, confident and human. And, crucially, interesting.
AI has given me new ways to research, organise and develop that work, but it has also raised the standard that finished content needs to meet.
I love what AI can do because I know what I want it to do. I also know when to close the chat window, switch off notifications… and write.
The businesses making the best use of AI will develop the same confidence. They will use it enthusiastically where it adds speed, scale and analytical power, and know when to put it to one side to protect the knowledge, judgement and personality that make their communication valuable.
There is no need to fear the machine. There is equally no need to worship it or put it in charge of the entire content department.
Just give it a desk, a carefully defined role and a human editor who knows exactly what good writing looks like.
About the author
What we do
Frontier15 helps brands achieve visibility in the era of AI search, creating content strategies that make your expertise discoverable and trusted by both people and machines. Get in touch and ask us how we can map a GEO-informed content strategy that helps future-proof your brand.