Stop Wasting Your Colleagues’ Time With Your Bad Writing

Bad internal writing is not a writing problem.

It is a respect problem.

Every time you send a vague, repetitive, generic, poorly reasoned document to your colleagues, you are asking them to do work you should have done first.

• They have to figure out what you meant.
• They have to separate facts from assumptions.
• They have to hunt for the actual decision.
• They have to ask follow-up questions.
• They have to sit through meetings that exist only because the document wasn’t clear enough.
• They have to reconstruct context from Slack, email, Jira, old conversations, and memory.

That is not collaboration. It is transferring your unfinished thinking to someone else.

AI didn’t create bad writing, but it changed who produces how much of it. There are two groups, and both got worse for your colleagues.

The people who used to write well but now lean on AI have gotten worse. The drafts come back polished, but the thinking underneath has thinned. They stopped doing the slow, uncomfortable work of figuring out what they actually meant, because the machine hands them something that sounds right fast enough to fool them.

The people who were already bad at writing are just as bad as before. But now they can produce far more of it, faster. So the volume of unclear, generic, unfinished thinking circulating inside your company has gone up sharply. That is what your colleagues feel every morning when they open their inbox.

The cost of producing words collapsed. The cost of reading them did not.

If you already know you’re creating slop and want help fixing it, I built a tool for that: AI Slop Check. Read this article first so you understand why the tool scores what it scores, then run your document through it before you send it to anyone.

The reader pays the cost

The reader pays the cost for your bad writing, not you.

Ten readers, fifteen minutes each, two and a half hours gone. That is what an unfinished document costs.

Twenty minutes of your time tightening before it goes out, or two and a half hours of theirs after it does. There is no version where the reader pays that is acceptable.

Bad writing creates hidden meetings

A bad document rarely dies quietly.

• It creates questions.
• Questions create replies.
• Replies create disagreements.
• Disagreements create calls.
• Calls produce slightly different recollections.
• Then somebody writes another document explaining what everyone thought the first document meant.

The meeting is often not the root problem. The root problem is that nobody wrote the issue clearly enough in the first place.

Bad writing hides weak understanding

If someone really understands a problem, they should be able to explain:

• What happened?
• What do we know?
• What do we not know?
• What evidence do we have?
• Why does it matter?
• What decision needs to be made?
• Who owns the next step?
• How will we know the issue is resolved?

If those answers are buried under vague language, generic recommendations, jargon, or three pages of background, there is a reasonable chance the writer hasn’t worked through the problem yet. The reader should not be forced to discover that on the writer’s behalf. I wrote the durable-artifact version of this list in If It Isn’t Written Down, It Didn’t Happen.

Bad writing weakens ownership

When someone writes the following, nobody owns anything.

• “We should improve communication.”
• “Engineering should follow best practices.”
• “Stakeholders need better alignment.”
• “The organization should implement stronger controls.”

• Who?
• What?
• By when?
• Based on what evidence?
• What changes?
• What counts as done?

A useful internal document makes ownership clearer. A bad one creates the appearance of action without committing anyone to anything. That is dangerous, because vague writing can make an unresolved issue look resolved.

Your colleagues are not your editors

Someone writes a rough thought. Sends it to five people. Then expects the group to “react.” Now five people are being used as an editing system.

Sometimes that is appropriate. Often it is laziness disguised as collaboration. Your colleagues should help improve difficult ideas. They should not have to clean up basic ambiguity, repetition, generic language, unsupported assertions, or obvious gaps that could have been caught before circulation. Their judgment is valuable. Do not waste it on work a machine could have flagged in thirty seconds.

Bad writing creates political problems

When the written record is weak, people can remember events differently. They can reinterpret decisions later. They can argue about what was intended. They can selectively quote conversations. They can change the story.

Good writing narrows that space. It establishes what was known, what was believed, what was decided, and what remains unresolved. That protects everyone.

It matters most when you are discussing failures, customer incidents, engineering concerns, operational problems, performance issues, or major decisions. The weaker the document, the more likely the organization falls back into anecdotes and personalities. Good writing keeps the discussion on facts. I wrote about the instrument that catches this in motion in Stop Telling People to Communicate Better. Instrument It.

Use the machine before using the humans

There is almost no reason today to circulate an important internal document without first having AI challenge it.

Not because AI should decide whether your thinking is correct. Because AI is cheap enough to catch obvious waste before expensive humans encounter it.

I built a simple tool for scoring internal business writing for unnecessary slop before it is circulated: AI Slop Check.

It asks a few questions about the document and the company, then builds a review prompt you paste into any assistant. The prompt scores the document from 0 to 10, where 0 is specific, evidenced, and grounded in your actual company and 10 is content that could have been written for almost any company with only the names changed. The central test is one sentence: could large parts of this document be dropped into another company’s document unchanged?

Treat the score as a gate before the document circulates. High score, keep working. Low score, the document has earned its readers. Nothing you enter is stored or sent.

Run the document through it first. If the score is bad, do not defend the document. Fix it.

Your colleagues have better things to do than finish your thinking for you.

If you don’t want to use the tool, the prompts at the bottom of this article give you the same ideas to run yourself. Use them. The point is not the tool. The point is that you attack your own document with a machine before a human has to.