Last week, someone on your team handed you a report. Polished. Well-formatted. Confident. You started reading, realized halfway down that it said nothing you could use, and you moved on. Then you spent your own time working out what the report was supposed to answer in the first place.
This problem now has a name. Workslop. Researchers at BetterUp and Stanford named it in 2025 to describe AI-generated work that looks like good work but lacks the substance to move a task forward. In the months since, most of the coverage has blamed the tool and prescribed the same fixes: better prompts, more AI training, tighter usage rules.
They’re treating a symptom.
Workslop isn’t a prompting problem. It’s a leadership problem. It’s what you get when nobody defined the outcome or the standard before the work started. The tool did exactly what it was told, which was nothing in particular.
What workslop actually is
The term came out of a September 2025 Harvard Business Review article from BetterUp Labs and Stanford’s Social Media Lab. Their definition is sharp: AI-generated content that masquerades as good work but lacks the substance to meaningfully advance a task. Think slick slides, a long report, tidy-looking code, all of it confident and none of it usable.
The numbers are worth sitting with. In their survey of 1,150 US workers, 40 percent had received workslop in the past month. Each instance took nearly two hours to sort out. They put the cost at $186 per employee per month, which runs to roughly $9 million a year for a company of 10,000 people. The receiver pays that tax, not the sender. The work looked done, so the burden just shifted downstream to whoever had to figure out it wasn’t.
You’ve seen the smaller version everywhere, including in your own building. The sixty-page report nobody reads. The tool someone stood up that scans, scores, and then sits, used by no one. The dashboard that measures activity and moves the business not one inch. None of that is a technology failure. It’s activity mistaken for progress, and AI made activity nearly free.
Why it’s a leadership problem, not a tooling one
Here’s the mechanism underneath it.
For seventy years, work has expanded to fill the time available. Cyril Northcote Parkinson named that in 1955. The law always had a ceiling, because time was finite. You only had so many hours, so eventually you had to stop and ship. AI removed the ceiling. Work still expands to fill the space available, but now the space is capacity, and capacity runs close to infinite. Nothing tells the tool it’s done, so it never is. It always has another draft, another alternative, another layer to go deeper.
Pull the deadline out and the deciding stops. That’s the soil workslop grows in. When no one has defined the outcome or the quality bar, capacity produces confident volume instead of useful work.
You can watch the same failure at company scale. Researchers surveyed roughly 6,000 executives across four countries. Eighty-nine percent reported no measurable productivity gain from AI, against more than $250 billion in corporate spend. That’s the signature of capacity expanding while value stays flat. It’s workslop measured in billions instead of slide decks.
The clearest way to see it is to picture a new hire. You hand a sharp new employee a project with no context, no example of what good looks like, and no sense of the lanes the work has to stay inside. They hand you back something polished and beside the point. You wouldn’t call that a talent problem. You’d recognize that you never briefed them. AI is the same, except it’s cheap and fast, so the failure scales in an afternoon instead of a quarter.
That’s the part the tooling story misses. AI will either amplify the effects of great leaders or expose their flaws. Workslop is the flaw, exposed. The gap between what a leader meant and what they actually defined now shows up at machine speed.
What this means for you
You can’t prompt your way out of workslop. You charter your way out.
Chartering is the move where you define the work before you open the tool, the same way you’d brief a capable person you trust. It comes down to four things.
Define the outcome, not the deliverable. The deliverable is the artifact, the report or the deck. The outcome is the effect it creates. “A board memo” is a deliverable. “A briefing that gets the board to approve the budget without a follow-up meeting” is an outcome. AI is excellent at deliverables and mediocre at outcomes, because it can’t read your mind about the effect you’re after.
Set the quality bar with examples, not adjectives. “Make it sharp” tells the tool nothing. Show it two pieces that hit the bar and one that missed, and name the difference in a sentence. That’s how you onboard a person, and it’s how you calibrate a tool.
Name the red lines. What is this work not allowed to be? The jargon you won’t accept, the assumptions you won’t let it make, the length it can’t exceed. Red lines have no exceptions.
Say what done looks like. Specific enough that a colleague could check the output against it and tell whether you hit the bar. If you can’t write down what done looks like, the work isn’t ready to start.
The companies pulling real value from AI do a version of this without thinking about it. They point the capacity at a defined outcome. The rest turn people loose with a powerful tool and mistake the motion for momentum. The discipline is the whole difference, and it’s a leadership discipline, not a technical one.
Where this holds, and where it doesn’t
Some workslop really is a skill gap. A person who’s never been shown how to direct a tool will produce more of it, and AI literacy helps at the margin. That part of the coverage isn’t wrong.
But for a senior team, literacy isn’t the root cause. Your people know how to write a prompt. What they were never given is a clear picture of what good looks like, because you were moving fast and the tool made it easy to skip that step. The fix isn’t another training. It’s the discipline of chartering the work before it starts.
And workslop is not an argument against AI. The gains are real and they’re coming. Erik Brynjolfsson, who tracks this closely, estimates US productivity grew about 2.7 percent in 2025, nearly double the average of the past decade. AI makes capable, well-directed people more productive. The dip most companies are sitting in right now isn’t the case against it. It’s the proof that direction is the variable. Accidentally great some of the time loses to intentionally good all of the time.
The teams harvesting the gains and the teams drowning in slop are using the same tools. The difference is leadership, not technology.
The question worth carrying into your next meeting
Are your people using AI to create business outcomes that matter, faster? Or are they just building more powerful ways to be busy?
The honest answer tells you whether you have a workslop problem, and it has almost nothing to do with which tools you bought. Clarity precedes momentum. That was true before any of this, and it’s truer now that the machine never stops on its own.
If you want a way to install the discipline, I wrote a short field guide on it. The Three Lenses Field Guide walks through how to charter AI work the way you’d brief your best people, and it includes the Charter template my own team runs on. You can download it here.
And if you’d rather talk through where the workslop is leaking into your organization, book a complimentary clarity call.
FAQs
Workslop is AI-generated work that looks polished but lacks the substance to move a task forward. A slick report, a clean deck, tidy code, all confident, none of it usable. Researchers at BetterUp and Stanford named it in 2025. The receiver pays the price, redoing work that only looked done.
Workslop comes from unchartered work, not weak prompts. When you open an AI tool without defining the outcome, the quality bar, or what done looks like, the tool fills the gap with confident volume. It hands you a deliverable when you needed an outcome. That gap is where slop lives.
You charter the work before you open the tool. Define the outcome you want, not the artifact. Show examples of good instead of describing it with adjectives. Name what the work can’t be. Say what done looks like. Ten minutes of that saves an hour of rework.
No. AI makes capable, well-directed people more productive, and the early gains are real. Workslop is what happens when capacity has no direction. The gap between the teams pulling value from AI and the teams buried in slop is a leadership gap, not a tooling one.
Researchers at BetterUp Labs and Stanford’s Social Media Lab named it in a September 2025 Harvard Business Review article. They defined it as AI-generated content that masquerades as good work but lacks the substance to advance a task. The label stuck because the experience was already everywhere.
