The report lands in your inbox looking finished. Clean headings. Confident tone. Charts that appear to mean something. You start reading and realize you’re doing the actual work now. The thing wasn’t thought through. It just looks like it was.
That’s workslop. AI-generated output that seems professional on the surface and fails the second someone has to use it. Someone else then spends the afternoon rewriting it, checking the numbers, and trying not to sound angry in the reply.
I’ve sat through briefings built this way. The slide deck is pretty. The person presenting it can’t answer the second question. You can see it in the room if you’re watching. Eyes drop to the notes. A small delay before the next sentence. Shoulders tighten. My Profitable Body Language book is blunt about incongruence. The mouth says the analysis is solid. The rest of the person is already looking for the exit.
Sun Tzu warned about mistaking appearance for strength. An army that looks ready and is already spent is a gift to the other side. He also warned about exhausting your own force on the wrong ground. Workslop does both at once. It creates the look of progress and burns the people who still care about the result.
The research on this is not subtle. Open-source surveys keep landing on the same pattern. A large share of people can name a specific piece of workslop that landed on them and cost real time. A lot of those same people admit they’ve sent it. Workers are spending a serious chunk of their AI time not creating anything, just fixing what the tool produced. Some of that time now has a name of its own. Botsitting. Hours spent verifying, rerunning, and cleaning up answers that sounded sure of themselves.
The ugly part is what it does to trust. When you receive slop from a colleague, you don’t just fix the file. You downgrade the person. Less capable. Less careful. Harder to work with next time. That’s not a software problem. That’s a command problem.
Most of this starts at the top. Leaders say “use AI” the way they used to say “be a team player.” No standard for what good looks like. No permission to say the draft is junk. Pressure to show volume. The people who still have standards become the cleanup crew. The people who don’t start signing off on whatever the machine spit out and calling it done.
I hit a version of this in my Redirecting Rejection book. You don’t beat bad material by white-knuckling through it. You change the frame. Stop treating every generated page as work. Treat it as raw terrain. Some of it is useful. A lot of it is noise dressed up as a deliverable. The second you accept the costume as the thing itself, you’ve already lost the morning.
So what do you do if you lead a team, or if you’re the one drowning in other people’s slop?
Stop measuring output by how finished it looks. Measure whether it moved the mission. A ten-page brief that creates six hours of rework is not a win. It’s a tax on the people who still know what good is.
Name the standard out loud. What does “done” mean in your shop. What has to be checked by a human before it leaves the room. If you won’t say it, people will fill the gap with speed.
Protect the people catching the slop. Middle managers are getting crushed right now. They’re supposed to coach AI use, validate the product, keep the timeline, and absorb the rewrite. Flatten the org on top of that and you didn’t get efficiency. You just removed the layer that used to stop garbage before it reached the decision.
If you’re producing the work, read it like an outsider before you hit send. If you can’t defend the second paragraph without looking back at the prompt, it isn’t ready. Sun Tzu didn’t send a force forward because the banners looked impressive. He sent it when the ground was actually held.
The shops that stay sharp won’t be the ones generating the most pages. They’ll be the ones who can still tell the difference between a finished thought and a costume.
Stay Strong, Stay Disciplined, Stay Dangerous,
–S