There’s been an influx of AI-written content across the internet over the past couple of years. It’s appearing on small business blogs, professional service websites and major publishing platforms, and much of it is technically competent.
The grammar is sound, the structure is clear, and the argument is usually presented in a careful, balanced way. The problem is that a growing amount of it sounds as though it came from the same author.
As people become more familiar with AI-generated writing, I think they’ll begin to recognise those patterns. They may not consciously identify the software involved, but they’ll notice when an article feels polished without appearing to come from anyone in particular.
That's a problem for a business blog because its purpose should extend beyond filling a page or attracting a visit from search. Getting the click is useful, but it only brings someone through the door. The content still has to convince that person that you understand their situation, have relevant experience and are worth listening to.
The reader is the person who ultimately matters. A search engine may introduce them to you, but it can’t make them trust you.
When a familiar voice disappears
I recently noticed a significant change in a client’s blog.
Their older articles had a recognisable voice. The writing was sometimes closer to a stream of consciousness than a carefully engineered content structure, but it felt genuine. The author understood the subject, showed empathy towards the reader and wrote about situations they had clearly encountered themselves.
The articles weren’t perfect, but they were believable. You could sense the person behind them.
The newer posts are very different. They use more single-sentence paragraphs, more deliberate attempts at impact and the familiar structure of an AI-generated article. Each side of the subject is treated carefully before the piece reaches a safe conclusion that few people could object to.
Personal details have been inserted here and there, presumably to restore some individuality. Unfortunately, they don’t. The personal sections sound like the original author, while the rest sounds as though it was produced elsewhere. The joins are visible.
That may be because AI wrote the first version, and the author later returned to add a few of their own experiences. Whatever the workflow, the result feels less trustworthy than the earlier, less polished writing.
Authenticity can’t be injected afterwards through a personal anecdote or a few informal phrases. The author’s experience needs to influence the article from the beginning: which part of the subject they choose to discuss, what they consider important, which advice they reject and where they eventually stand.
A regular reader may not declare that the writing is AI-generated. They may simply feel that the person they were used to hearing from has disappeared.
For a business whose customers are buying expertise, reassurance or personal support, that change is commercially important.
Trust has always been difficult online
Anyone can make claims on a website. A business can describe itself as experienced, caring, innovative or customer-focused, but the visitor has no obligation to believe any of it. They are only words written by the organisation making the claim.
A useful blog can provide better evidence. The reader can see whether the author understands the problem, whether they recognise its practical consequences and whether their advice appears to come from experience.
That is particularly important where the service cannot be assessed fully before it is bought. A prospective client choosing a consultant, therapist, solicitor, specialist tradesperson or other professional service is partly judging the people behind the offer. They want some indication of how those people think and whether they understand the situation they’re being asked to deal with.
Good writing can provide that indication, while generic writing tends to obscure it.
This is why I think heavily generated content presents a trust problem. It can sound credible without demonstrating that the person publishing it has any direct knowledge of the subject. The reader receives a competent explanation, but learns very little about the business responsible for it.
If the same article could appear on a competitor’s website with little more than a change of logo, it isn’t contributing much to the reader’s decision.
Experience gives you something your competitors don’t have
An authentic voice is sometimes treated as a matter of style. Advice about “humanising” AI content often suggests adding jokes, contractions, informal language or the occasional personal story. While those devices can alter the tone, they don’t create authenticity.
The most important thing an author brings is experience. You have been exposed to particular clients, problems, mistakes and outcomes. Those experiences influence what you notice and how you explain it.
A general article might accurately describe a common problem. Someone who has dealt with it repeatedly is more likely to mention the awkward practical detail, the assumption that usually turns out to be wrong or the reason an apparently sensible solution fails in practice. Readers with the same problem recognise those details, and can tell when the author has been there before.
AI can help describe an experience that you provide, but it has no experience of its own. It hasn’t worked with your customers, watched a project go wrong or had to deal with the consequences of poor advice. It can reason from the information available to it, but it can’t supply your commercial history.
And there's the crux. Your competitors may be able to publish on the same subjects; they can repeat the commonly accepted advice and cover all the expected points. What they can’t honestly reproduce is your experience of those situations and the conclusions you’ve drawn from them; that produces individual, trustworthy content.
A point of view makes the difference visible
Experience on its own can produce useful material, but it's judgement that gives it direction. Two people can encounter similar situations and reach different conclusions, but what you choose to recommend, challenge or dismiss reveals far more than a neutral explanation of the subject.
That point of view may be contentious. I don’t see a problem with that, provided it is informed, fair and something you’re prepared to stand behind. Business writing doesn’t need to behave like an impartial reference work. If you believe a common approach wastes money, say so and explain why. If your experience suggests that a fashionable solution is unsuitable for most of your clients, that's worth publishing. The reader will decide whether your judgement makes sense to them.
This is one of the areas where AI-generated writing often becomes noticeably weak. It is very good at producing credible, grammatically correct content that addresses all sides of an argument. The conclusion is usually safe, reasonable and difficult to dispute, but it's rarely noteworthy.
AI can be prompted to argue almost any position; however, the author must decide what they genuinely believe and why. If AI supplies the argument, the examples and the conclusion, the finished article may be competent, but it won’t tell the reader much about the person whose name appears above it.
Search visibility doesn’t complete the job
Businesses often publish articles to appear in search results, which is a reasonable objective. A useful article can attract people actively looking for help on the subject it covers. However, appearing in the results and earning the click are only the first stages.
Once the reader arrives, the article must justify their attention. It needs to give them something beyond the same general explanation they could have found elsewhere. That may be a specific example, a clearer interpretation of the problem or an experienced view on what they should do next.
A page can rank and still achieve very little commercially. If the visitor reads a generic article, learns nothing about the business and has no reason to remember who published it, the traffic may have limited value. The aim isn’t to make generated content look human enough to satisfy an algorithm; it’s to reward the person who took the time to read it.
Search is the route by which that person may find you, but the content is what influences what happens next. Never lose sight of that.
Where AI is genuinely useful
None of this means I think AI should be excluded from the writing process. I’ve tried many different workflows for my own work and to understand what might help my clients. It can save a considerable amount of time when used at the right stage of the process.
It is particularly useful for brainstorming. Talking through an idea with AI can help identify the strongest part of the subject, reveal gaps in the argument and prevent the article from wandering into areas that don’t belong. It can also accelerate research, although claims still need to be checked against reliable sources. AI can help locate relevant information, compare different positions and make a large amount of material easier to understand.
At the editing stage, it is good at finding grammatical errors, repetition and contradictions. People often work out what they think while they are writing, so it isn’t unusual for a draft to begin in one place and finish somewhere slightly different. AI can point out where the reasoning no longer lines up without deciding which position the author should take.
A provisional structure can also be useful, provided the writer doesn’t then fill in each heading mechanically. A framework should help organise the thinking. When it becomes too visible in the finished article, the writing starts to feel assembled.
The most useful approach I’ve found so far is dictation. Talking through the subject captures my argument, examples and natural way of explaining the problem. The transcript is repetitive and untidy, but it contains the things that a generic instruction to “write a blog post” does not.
AI can then organise that material, remove obvious repetition and help turn it into readable prose. The danger is that it may also smooth away the judgement and irregularity that made the source material authentic in the first place, so the author still needs to review the result carefully.
Dictation doesn’t create authenticity by itself, but it does give the AI authentic material to work from.
Give the reader something worth finding
So, where do I stand?
A business blog should show more than knowledge of a subject. It should give the reader some sense of the people behind the business, how they approach problems and whether their judgement can be trusted.
AI can help with research, structure and editing. It can challenge an argument and make the writing clearer. Those are valuable uses, and they can reduce the time needed to turn an idea into a finished article. You need to remember that the experience, examples, and point of view still have to come from the author.
Before publishing, ask whether the article could have appeared on any of your competitors’ websites. If it could, it might be accurate and well-written, but is it worthwhile? What could you bring to the argument no one else could?