Workslop is a training problem, not a tool problem

A Stanford Social Media Lab / BetterUp Labs study published in Harvard Business Review (Sept 2025) put a name on the thing everyone was feeling: “workslop” — AI-generated work that looks polished but lacks substance. 40% of 1,150 US full-time employees said they'd received it in the past month, and the researchers estimated about $9M a year in lost productivity for a 10,000-person company. Scale that down to a 50-person firm and it's still someone's whole week, every month, quietly redone.

Why the slop happens

Not because the model is bad — because the ask was weak. “Write a client update” without the context, the audience, the constraints, or a review step produces exactly the polished-hollow output the study describes. The skill missing isn't typing; it's provisioning context, choosing the right task for AI, and verifying before sending.

Those are learnable. They're the first three of our eight practice areas (context provision, task–AI fit, attribution & citation), coached in the moment of work rather than in a course nobody finishes — 35% of employees report no AI training at all, and 85% say they can't apply the training they did get to their daily work (Study.com, State of AI Jobs and Skills 2026).

What this means for owners

Every workslop incident is a coaching opportunity your current stack throws away. Catch the weak draft at the point of creation, teach the fix in the flow of work, and the slop rate drops while the adoption curve rises. That's the “destroy the AI-slop stigma” mission: great AI use is a trained skill, not a talent.