What happens when someone lets AI generate a piece of work and skips their own review and editing role, passing the raw or barely touched output on to a colleague, manager, or client as if it were finished.
Also known as AI workslop
Workslop is what happens when the human in the loop role gets skipped. Someone asks an AI tool to draft a deck, a report, or a piece of code, and instead of carefully reviewing, editing, and taking ownership of that output, they pass it along mostly as is, often asking someone else to give feedback or catch what they should have caught themselves. The term was coined by researchers at BetterUp Labs and Stanford’s Social Media Lab in a 2025 study, and describes work that masquerades as finished but lacks the substance to actually move a task forward.
As Louis Carter, CEO of the Best Practice Institute, put it in coverage of the study: the reason it is slop is because the human element, the human in the loop, is not implemented enough for the actual output. The AI did its part, the person just did not do theirs. In the study’s numbers, 40 percent of desk workers said they had received workslop from a colleague in the past month, and each instance cost the receiver close to two hours to sort out, an average of $186 per employee per month once it adds up.
Agencies run on drafts moving between people under deadline pressure, which makes it easy for the human in the loop step to get quietly skipped and replaced with a note that just says thoughts.
It shows up as a shortcut, not a shortfall in skill. A strategist who is perfectly capable of reviewing an AI drafted competitive analysis skips that step under deadline pressure and forwards it as is, asking a colleague to weigh in instead of doing the first pass themselves. The AI did not fail, the review step did.
It quietly reassigns work instead of eliminating it. The time saved by not reviewing the AI’s output does not disappear, it gets pushed onto whoever receives the deck, the brief, or the script next, and they usually spend more time untangling it than the original author would have spent getting it right the first time.
It erodes trust inside the team fast. Once a colleague has been handed workslop by the same person twice, they start reviewing everything that person sends more skeptically, which slows the whole team down even on work that was actually careful.
A senior strategist needs a first draft pitch deck for a new business meeting in two days. They generate the full deck in a tool like Gamma from a short prompt, skim it once, and send it to a creative director with a one line message: rough draft, let me know what you think. The creative director opens it expecting a starting point to react to, but finds generic positioning language, a competitor slide with outdated logos, and a strategy section that does not actually reflect anything specific about the client’s brief. Instead of giving quick feedback, the creative director ends up rewriting half the deck themselves the night before the pitch. Afterward, the team agrees on a simple standard: whoever generates a first draft with AI owns a real editing pass before it goes to anyone else, checking it against the actual client brief, not just asking a colleague to find the problems for them.
The automations and agents module of the workshop teaches you how to build AI workflows that compress the busywork without taking the craft out of the studio.