The failures are not random. They cluster around a few consistent underlying problems, and once you can see the pattern you cannot unsee it. Get to the root of each cluster and you have a quality filter you can apply in about thirty seconds to anything you or anyone else writes about you.

Words That Have Done Too Much Time
Start at the surface, because the surface is where most bios go wrong first.
“Passionate.” “Boundary-breaking.” “Genre-defying.” “Eclectic.” “Unique.” “Visceral.” “Emerging.” “Up-and-coming.” “Quickly establishing herself as a sought-after.” “One of the most.” Read enough bios and these phrases stop registering as words at all. They are placeholders dressed as descriptions, and experienced readers, the promoters, journalists, and label people whose attention you are trying to earn, process them as noise. They tell the reader how you want them to feel without giving them a single reason to feel it.

The test from Part 3 applies at word level, not just sentence level. If a phrase could sit unchanged in any artist’s bio regardless of what that artist actually sounds like, it is doing no work. “Passionate about music” has never once persuaded a booker to listen to a mix. It is the verbal equivalent of a shrug.

Genre-label overload is the next variant of the same problem. The instinct behind it is understandable: you play a lot of different things, you do not want to be boxed in, you want the bio to cast as wide a net as possible. The effect is the opposite. “DJ X spins everything from the deepest progressive electro-house to the hardest broken-beat trance” tells an experienced reader exactly one thing: that the artist has not yet found the language to describe what they actually do. Stacking genre labels is not the same as communicating range. The Texture Exercise exists precisely because specific atmosphere and feeling do more work than any number of compound genre names stacked against each other.

The “X meets Y” comparison in a full bio is a subtler trap, because Part 3 acknowledged that comparisons can be useful at the one-liner stage as compressed shorthand for a sound that sits at a real and identifiable intersection. In a full bio they do real damage. They anchor you to references that may have nothing to do with what you are making in two years. They substitute someone else’s identity for yours at the exact moment when yours most needs to be established. And they read, to anyone who has seen a few hundred bios, as a sign that the artist ran out of their own words and reached for borrowed ones.

The Right Words, the Wrong Story

These failures go a level deeper. The problem is not the choice of words. It is the choice of material.

The non-story bio is easy to recognise and harder to write your way out of than it looks. “A DJ and producer who has been passionate about electronic music since childhood and believes music has the power to bring people together.” Everything in that sentence is technically true of most artists working today. It has no point of view, no distinguishing detail, and no reason for anyone to keep reading. It does not draw the artist out from the background; it dissolves them into it. The problem is not bad writing. It is that nothing specific enough to be interesting has been put on the page.

Irrelevant background detail is related but different. This is the bio that tells you the artist’s mother played piano, that they studied something unrelated before discovering electronic music, that they grew up listening to whatever was on the radio. Any of these might be genuinely interesting as personal history. None of them belongs in a bio unless there is a direct and legible line between that detail and the music the artist makes today. The Origin Excavation in Part 2 was not an invitation to include everything from your past; it was a process for identifying which past details actually connect to your present sound. The bio is not a CV. Every sentence should earn its place by illuminating the music, and the four words you identified in Part 2 are the sharpest available test: if a detail is not consistent with those four words, it probably does not belong here.

Some Bios Lie
These are not craft failures. They are decisions, and the reader can usually tell.

Dishonesty in a bio takes a few forms. Venues you have not played, credits you have not earned, a residency that was actually one night, a collaboration that was a file swap with someone you have never met. The South African electronic music scene is smaller than most people think, and reputations, in both directions, travel faster than most people expect. This is not a morality argument. It is a practical observation about how things actually work. The bio you submit is not just a document; it is the beginning of a professional relationship, and the relationship starts with whatever you decide to put in it.

The AI tell is a failure of a different order and it deserves more space than the others, not because it is more common but because it is more nuanced, and because it sits at the centre of something the whole series has been building toward.

An unedited AI-generated bio has a recognisable signature. It is not that it is badly written in the conventional sense; it is often grammatically clean and structurally coherent. The problem is that it is flat in a specific way: the expected phrases appear in the expected order, the arc moves from introduction to credibility to aspiration exactly as a template would, and there is no specific texture anywhere, no detail that could only have come from someone who has actually lived this. The sound description could describe almost anyone. The credibility signals are present but feel generic. The whole thing reads like a form that has been filled in rather than a portrait that has been painted.

Promoters who open two hundred EPKs a month, journalists who have profiled dozens of artists, label people whose inbox is permanently full: they identify this quickly, not always consciously, but the response it generates is a specific one. Mild interest that does not build into anything. The bio does not fail dramatically. It simply does not do the work it needs to do, which is to make a person who does not know you yet want to know more.

The solution is not to avoid AI tools. The solution, which is what Part 5 is entirely about, is to do the work in Parts 1 through 4 first, so that what you bring to the tool is specific enough that the output cannot default to the generic. The AI cannot invent your texture words, your origin detail, your four words, your specific credibility signals. Only you have those. Give them to the tool and the output changes completely. Skip them and you get the Hall of Shame entry you were trying to avoid.

Exercise 5: The Annotation (Continuing from Exercise 4 in Part 3)

You already know Siya Dlamini from Part 3. You know his one-liner. You know his short bio opening sentence. Now read a different version of him:

Durban-based DJ and producer Siya Dlamini is a passionate and boundary-breaking artist who has been pushing the limits of electronic music for years. Growing up in KwaZulu-Natal in a household filled with music, Siya’s journey began at a young age and led him through a diverse range of sounds. Often described as where gqom meets deep house meets afro-tech and beyond, his unique and distinctive sound defies categorisation and has quickly established him as one of the most exciting emerging artists on the South African circuit. He has performed at clubs across Johannesburg, Cape Town and Durban and released music on several independent labels. With a visceral and genre-defying approach to the dance floor, his upcoming work promises to showcase his versatility and his deep commitment to electronic music.
Same artist. Before the work in this series was done.

Annotate it sentence by sentence. Name the failure mode operating in each one, using the three clusters from this piece: words that have done too much time, the right words but the wrong story, or the lie, and the specific variant within that cluster. When you are done, take the opening two sentences and rewrite them using the Part 3 system: specific texture language from the Texture Exercise, one identifiable origin detail, no borrowed phrases. Use the good version of Siya from Part 3 as your benchmark, not your ceiling. The rewrite does not need to be finished. It needs to be more specific than what is there. That is the only test.

The Pattern, Once You See It
Every failure in this piece has the same root: the work in Parts 1 through 3 was not done, or was not done far enough. The clichés arrive when the texture language has not been pushed to the point of genuine specificity. The structural failures arrive when the Origin Excavation has not been filtered properly. The AI tell is the clichés and structural failures produced at scale and speed, with no lived detail anywhere in the output to interrupt the template.

The filter is already in your hands. You built it.

In Part 5, we put it to work.

Next: Part 5: Using AI the Right Way – How to Brief, Prompt and Evaluate an LLM to Write Your Bio

Credit: JP Vorster – Beatlab Editor

ARTICLES