You were hired to build campaigns. Somewhere along the way, the job became something else entirely.
A recent Stanford and BetterUp study surveyed more than 1,100 US desk workers and found that 40 percent of them say AI saves them no time at all. Meanwhile, 92 percent of senior executives say AI makes them more productive. That gap tells you something important about where AI's costs actually land. They don't land on the people approving the tools. They land on the people using them.
Marketing managers at growth-stage companies describe the same shift. The tools arrived, the AI access expanded, and instead of producing more campaigns with less effort, the work multiplied in ways nobody planned for. Drafts come faster, sure, but everything that happens after the draft takes longer.
The Labor That Wasn't in the Job Description
The pitch for AI in marketing was straightforward. Generate first drafts faster. Produce more content per cycle. Free up your team for strategic work. And the first part of that pitch delivered. First drafts do come faster. Sometimes they arrive in seconds.
What nobody accounted for was the labor that follows. The Stanford/BetterUp researchers coined a term for it:
Workslop.
It's the flawed, superficially polished AI output that requires heavy editing, fact-checking, tone correction, and often a complete rewrite before it can be used. Their study found that 40 percent of workers encountered workslop within a single month.
For marketing teams, this plays out in a specific pattern. The AI produces a draft of a campaign email. The draft sounds confident. It hits the right length. It even mirrors the brand voice, mostly. Then someone reads it closely and finds a proof point that doesn't exist, a CTA that contradicts the campaign's actual offer, and three sentences that say the same thing in slightly different ways. Fixing all of that takes longer than writing the email from scratch would have.
That's the second job. The one nobody hired you for.
Why the Editing Takes Longer Than the Writing
Editing AI output is different from editing human output. When a colleague writes a draft, you can see where their thinking broke down. You can trace the logic, identify the weak spot, and strengthen it. The rest of the draft usually holds.
AI output doesn't work that way. The confidence is uniform. Every sentence reads like a final draft, which means you can't skim for weak spots. You have to read every line with the same level of scrutiny because the errors hide behind fluent prose. A fabricated statistic sits in the same paragraph as a verified one, delivered with identical conviction.
"Quality decreased significantly, time to produce a piece of content increased significantly and, most importantly, morale decreased. Everything got a whole lot worse once they rolled out AI."
That's from a copywriter at a cybersecurity firm, speaking to The Guardian about what happened after his company mandated AI tools. The marketing manager reading this already knows the feeling: a draft that looked ready at 9am turned into a four-hour editing session by noon, and the campaign brief that was supposed to ship moved to tomorrow.
The problem compounds when multiple people on a team are each generating AI drafts. Now you're reconciling work that came from different prompts, different sessions, and different interpretations of the brief. The researchers at Stanford found that workers spent an average of 3.4 hours per month dealing with workslop. On a marketing team running two or three campaigns simultaneously, that number understates the reality.
The Cost Nobody Is Tracking
Here's what makes this expensive. The rework labor is invisible. It doesn't show up in any dashboard. There's no line item for "hours spent fixing AI drafts" or "campaigns delayed because three people's AI outputs contradicted each other." The work just absorbs into the day, displacing the strategic work that was supposed to benefit from AI in the first place.
For a five-person marketing team, there's no slack in the schedule. Every hour spent reconciling AI output is an hour that didn't go toward campaign strategy, audience research, or the creative decisions that actually move pipeline. The Stanford/BetterUp team estimated that workslop costs a 10,000-person organization $8.1 million in lost productivity. Scale the dollar figure down and it shrinks. The operational impact doesn't.
The executives don't feel this. Remember: 92 percent of them say AI makes them more productive. They're seeing the speed of the first draft. They're not seeing the three rounds of editing, the cross-checking against approved proof points, the tone realignment, and the quiet frustration of a marketing manager who used to spend their time building campaigns and now spends it cleaning up after a chatbot.
Where the Time Actually Goes
If you're a marketing manager at a B2B company right now, do an honest audit of your last two weeks. Count the hours you spent on each of these: writing prompts and iterating on AI outputs until they were usable, editing AI drafts for accuracy, tone, and consistency with your actual brand, reconciling work that multiple people or tools produced independently, and explaining to someone upstream why the "AI-generated first draft" isn't ready to publish.
Now count the hours you spent on this: defining campaign strategy, mapping content to buyer segments, building a conversion path that connects the pieces, and reviewing performance to inform the next cycle.
If the first list is longer than the second, you've found the workslop in your system. The one that appeared when the AI tools arrived and nobody changed the operating system around them.
Speed without structure doesn't scale the output. It scales the cleanup. And the cleanup has a cost that compounds every cycle, even when nobody is measuring it.
The teams pulling ahead aren't the ones with better prompts. They're the ones who built a system that controls what the AI produces before anyone starts typing.