Researchers at Stanford and BetterUp have a name for it: workslop. It's AI-generated work that "appears polished but lacks real substance."
Forty percent of the 1,150 US desk workers they surveyed had received some in the past month. Each piece took almost two hours to untangle, costing about $186 per employee per month. Half the recipients thought less of the person who sent it. (Axios, on the study published in Harvard Business Review)
The bigger picture isn't better. MIT's Project NANDA found that about 95% of organizations see no measurable return from generative AI, despite $30–40 billion in enterprise spending. (The GenAI Divide, via Virtualization Review)
The easy conclusion is that AI makes junk. It's the wrong one.
The job moved up a level
Slop is what you get when execution has no architecture behind it.
Here's the uncomfortable part. Some slop is self-defense. Writers, designers and developers keep producing the volume they always did, because volume was the job. If the machine can make the output, the thinking goes, I'd better keep making it too.
That instinct is understandable. It's also aimed at the wrong target. The work didn't disappear. It moved up a level, from making the thing to designing the system that makes the thing. AI handles the execution. People own the structure and the strategy.
Three frictions, three kinds of slop
Slop isn't random. It comes from the same three frictions that slow every midmarket company. AI just makes them louder.
1. Strategy Friction: fast output in no direction. When nobody can say what the company stands for, AI fills the gap with something generic. A thousand on-brand-ish posts don't add up to a position. The fix isn't more prompting. It's a clear, simple strategy the whole organization can act on, and a voice someone actually owns.
2. Execution Friction: tools bolted onto old workflows. MIT's researchers found that most AI pilots stall because the tools don't fit the way work really gets done. They don't learn, and they sit outside the workflow. Speed only counts when AI is built into the work itself, inside rules someone designed on purpose.
3. Value Friction: time saved, nothing created. This is workslop's signature. It looks finished but adds nothing new. Cutting hours is table stakes. The point is to create what you couldn't do before: a proactive pricing review instead of a reactive one, or a customer review that uncovers millions in savings instead of reporting spend.
The new job descriptions
You still need writers, designers and developers. You need them for the architecture, not the volume.
Designer. Old job: make every asset. New job: create one ownable brand system, then let AI produce the rest on brand.
Writer. Old job: write every draft. New job: own the strategy, the narrative and the voice rules AI writes inside.
Developer. Old job: write every line. New job: engineer the contracts, architecture and governance AI works within.
Leader. Old job: approve the output. New job: set direction, decide what AI should create, and keep it curated.
The common thread is control. A designer who builds the brand system controls every deck that follows. A developer who writes the governance controls every agent that runs. That isn't a smaller job. It's a bigger one.
Architects, not operators
At Velocity we call this work architecture, and we do it with your people, not to them. We co-build the strategy, so your AI has Direction. We build AI into the workflow, so you get Speed. And we design for what's newly possible, so you get Value. Everything stays under your control.
The companies that win won't be the ones that produce the most AI output. They'll be the ones whose people know their jobs have changed, and build the architecture that makes every output count.
Consultants fix problems. Velocity Architects install momentum you control.
Lead with Velocity. Start with a free Velocity Diagnostic: we'll find your three frictions and show where AI can create value you couldn't before.
See it at work: What we do and Results at BGR.
Sources: Axios on the Stanford and BetterUp workslop study (Harvard Business Review, Sept 2025) · MIT Project NANDA, The GenAI Divide (July 2025), via Virtualization Review
