The output standard

AI is very good at producing work that looks finished: a tidy email, a confident summary, a long report. Sometimes it is finished. Sometimes the real work has only moved, from the person who made it to the person who receives it. The output standard is four rules that stop that happening. They apply to anything AI helped you make, before it goes to someone else.

Workslop

In September 2025, researchers at BetterUp Labs and the Stanford Social Media Lab gave the problem a name in Harvard Business Review. They called it workslop: “AI generated work content that masquerades as good work, but lacks the substance to meaningfully advance a given task.” Then they asked 1,150 full-time US employees how often it reached them.

The sender saves a few minutes. The receiver has to work out what's wrong, check what's true, cut what isn't needed, and often go back to ask what was meant. The time AI saved gets spent again, by someone else, and the sender never sees it happen.

It isn't only a problem with new starters, either. In the same survey, workslop travelled in every direction.

Why it happens

Nobody sets out to send workslop. It happens for ordinary reasons. When people feel pushed to use AI, or praised for how much they produce, checking is the step that gets squeezed. When nobody has said what “done” means, “it looks finished” becomes the test. And because the cost lands on someone else, the sender never sees the hour and fifty-six minutes it took to sort out. The output standard deals with all three: it says what done means, and it keeps the cost with the person who saved the time.

The four rules

• Labelled: say where AI helped. Not as a confession, as information. The reader knows what to look at more closely, and nobody is surprised later. One line is enough.
• Owned: your name is on it, and you can explain it. If you can't say why a number is what it is, or why one option beats another, it isn't ready. “The tool said so” is never the answer.
• Short: no longer than the reader needs. AI makes length free for the writer and expensive for the reader. A summary longer than the thing it summarises is a cost, not a help.
• Checked: the facts that matter are checked. Not every word. The things that would cause harm if they were wrong: numbers, names, dates, quotes and anything you promise.

Try it on six outputs from an ordinary week. Pick every rule each one breaks, then see who ends up paying.

Your label line

Labelling only helps if it's quick and consistent, so many organisations set one standard line. If yours has, use it exactly as written. If not, build your own below and ask whether there's a standard you should be using. Your line goes on your completion record.

A thirty-second habit

You don't need a process for this. Before you send anything AI helped make, ask yourself four questions:

1. Have I said where AI helped?
2. Could I explain every part of this if someone asked?
3. Is it as short as the reader needs?
4. Have I checked the facts that would matter if they were wrong?

If any answer is no, fix that one thing first. Usually it takes a few minutes, and those minutes are the difference between saving time and moving it onto someone else. Put your label line where you'll actually use it: your email signature, your document templates, the end of your reports.

Your label line is saved to your completion record. Next: what to do when you want to change something bigger than your own work.