# Why I still write my own drafts

URL: https://ekom.blog/why-i-still-write-my-own-drafts/
Published: 2024-12-31 · Updated: 2026-11-07
Author: Ekom
Topics: AI & search

> Why I still write my own drafts

A blank document can feel like the hardest part of writing. It is tempting to ask a tool for a finished article and publish whatever comes back. I still write my own drafts because the first version is where I decide what I actually think, what I know and what I need to verify. That work cannot be delegated to a fluent paragraph generator.

## A draft is a thinking tool

When I write a draft, I have to choose the point of the article. What problem does the reader have? Which answer is most useful? What evidence supports it? Which detail is uncertain? Those questions often change the structure before the piece is polished. If I skip them, a smooth introduction can conceal that I have not decided what the article is about.

A draft also reveals gaps. I may know the general principle but not a specific client result, date or platform limitation. That is the moment to research, ask for the fact or leave it out. A made-up number can make a paragraph sound persuasive while making the whole post less trustworthy. Better to mark a missing detail than to invent one.

## Use tools as assistants, not witnesses

AI tools can help brainstorm questions, organise notes, suggest an outline or point out a confusing sentence. They can also return outdated facts, flatten a careful distinction or produce a confident claim with no reliable source. I treat the output as a prompt to inspect, not as evidence that something happened.

That distinction matters in SEO writing. Platform interfaces change. Google’s terminology changes. Search results differ by location and query. An answer that sounds plausible may describe a feature that was never available to the audience in question. Verify the detail against a reliable source and date the claim where it matters. If I cannot verify it, I should not write it as fact.

## Keep the human judgement visible

A personal blog should sound like its author, not like a collection of generic paragraphs. That does not mean every article needs a dramatic story. It means making deliberate choices: which example is useful, which caveat belongs beside a recommendation and where the advice stops applying. The author is responsible for the judgement even when a tool helped with editing.

I also want to be careful with experience. Working on client sites teaches me things, but it does not give me permission to name clients or claim results I cannot share. A lesson can be explained without exposing a business. If a post needs a specific metric or anecdote that only I can supply, I should add it myself or leave the claim out. Privacy and accuracy matter more than a colourful case study.

## Edit for the reader

After the first draft, I check whether the opening names a real problem, whether the answer arrives early, and whether each section earns its place. I look for repeated points, unexplained acronyms and advice that ends without a next step. Then I read the article aloud or scan it as someone with limited time would. If the reader has to work too hard to find the answer, the structure needs another pass.

Tools can help with grammar, but they cannot decide whether the article is fair, useful or appropriate for a particular audience. They do not know which personal detail should stay private. They cannot take responsibility if a recommendation misleads someone. That responsibility remains with the person who publishes the work.

## Publish only what I can stand behind

A good draft is not necessarily a perfect draft. It is one whose claims I can explain and whose advice I believe is useful. Some posts need a source link. Some need a caveat. Some need me to check my own notes before publication. I would rather make that uncertainty visible than hide it under polished wording.

AI is part of how people work now, and I expect to keep using it where it saves time without weakening the thinking. The boundary is simple: I own the facts, the judgement, the voice and the final decision. The tool can assist, but it cannot be the witness for my experience.

That is why I still write my own drafts.

## Keep the process honest when AI helps

If I use a tool to outline a post or suggest a clearer heading, I still need to verify every factual statement and revise the language so it fits the argument. I should not ask a model to invent a client anecdote, estimate a result or supply a personal opinion I do not hold. A fluent paragraph can make uncertainty disappear from view, which is exactly when a writer needs to slow down. I want to keep source notes and distinguish verified facts from interpretation. When a detail needs my own confirmation, I mark the gap and check it before I publish. This is not about refusing useful assistance; it is about staying accountable for the finished work. Readers see the author’s name, so the author must be able to stand behind the claims.

## Keep the voice accountable

When a sentence says “I found” or “I learned”, it should describe something I can genuinely explain. If I am summarising a public announcement, I should attribute it or link to the source rather than make it sound like personal experience. If I am giving an opinion, the reasoning should be visible. This separation makes writing clearer: readers can distinguish what happened, what I believe and what they might try on their own site. It also keeps a personal blog personal without turning every practical guide into an unsupported story.

What part of your writing process would improve if you used AI to support the work while keeping the judgement yours?

## Related reading

- AI and SEO in 2024: what actually changed

- Why 2024 rewarded brands, and what small sites can do

---

From Ekom (https://ekom.blog/). More: https://ekom.blog/llms.txt
