Better prompts can’t solve your workslop issue
▼ Summary
– A marketing team’s efforts (prompt library, voice guide, training, office hours) failed to eliminate “workslop” — low-quality AI-generated content like half-finished briefs and off-target copy.
– Research shows 40% of employees received workslop in the last month, costing roughly $9 million annually at a 10,000-person company, while only 19% of workers have clarity on what AI work is appropriate.
– The root cause is a coordination-of-learning problem: individual team members learn effective AI practices in isolation, but that knowledge never transfers across the team, leading to persistent poor output.
– The solution is an “AI activation hub” — a small group that actively moves learning between people by atomizing insights, pairing members, maintaining a searchable knowledge engine, and measuring AI’s impact.
– Marketing teams that solve workslop will build this connective layer so that when one marketer learns something, the rest of the team uses it within a week, with efficiency following as a result.
On a recent strategy call with a mid-sized B2B SaaS marketing leader, I heard a familiar story. Her team had tried everything to eliminate their workslop problem. They built a shared prompt library in Notion. They published a refined brand voice guide. They ran two rounds of AI literacy training. They even held monthly office hours where the most experienced AI users mentored the rest. The CMO personally drafted a memo modeling thoughtful AI use, gently nudging the team toward substance over volume.
None of it worked. Workslop kept piling up. Half-baked briefs that read like rough drafts. Slide decks that looked polished but collapsed by the third bullet. Newsletter copy that met the brief but missed the audience entirely.
Workslop is easy to name but harder to diagnose. Its root cause lives deeper in the organization than most leaders are willing to dig.
Where the workslop conversation currently stalls
Data from BetterUp Labs, Stanford, and two HBR studies (September 2025 and January 2026) paints a stark picture. Forty percent of employees received workslop in the last month. Each instance cost nearly two hours to clean up. For a 10,000-person company, that adds up to roughly $9 million annually, wasted on fixing AI-generated work that was supposed to save time.
The most telling number comes from Asana’s State of AI at Work research. Only 19% of knowledge workers say they have clarity on what types of work AI should handle in their role. That single statistic explains the rest.
The dominant fix right now mirrors what my call-mate has been running for six months. Leaders should model purposeful AI use. Teams should set clear guardrails. Individuals should develop what BetterUp calls a pilot mindset. Human-AI output should meet the same standard as human-only output. Greg Kihlstrom’s recent MarTech piece pushes this further, urging marketing leaders to define handoff lines with IT, legal, and procurement.
All of this is correct. None of it is wrong. But every solution places the burden on the same spot: the individual prompter, the individual leader, the individual mindset. That’s the layer most teams have pulled at for the last 18 months. The real fix sits somewhere else entirely.
Where the system actually breaks
Workslop happens when individuals produce AI-generated work with no connective tissue between them.
On a healthy marketing team, learning must move fast. The content specialist figures out in week one that this model needs a longer brief and a tighter persona. The designer discovers in week two that the image tool wants brand colors in hex, not plain English. The email marketer learns in week three that AI subject lines sound generic unless you feed it the last three high-performing ones.
Each of those insights is real. Each one was earned. But in most marketing teams I see, none of that knowledge travels far. The content specialist doesn’t know what the designer figured out. The email marketer doesn’t know what the content specialist learned. There’s no shared space where someone says, “This worked, here’s how I got there, try it your way and tell me what improves.”
What you end up with is a team of skilled individuals, each running their own little R&D project in parallel. Each one gets better in their slice. But the team’s combined output still produces workslop because no individual’s learning ever reaches the next person. When someone moves teams, that learning walks out the door with them.
This is the part of the workslop problem nobody names clearly. It’s a coordination-of-learning problem. It doesn’t get solved by another training session or a sharper brand voice guide. It gets solved by building infrastructure that carries learning between people.
What the AI coordination fix looks like
In my book “Hyperadaptive,” I call this connective layer the AI activation hub. A hub is a small group of people inside the organization, virtual, physical, or both, whose job is to keep AI capability flowing through the rest of the team in both directions.
It’s different from a help desk, a prompt library, or an AI ticketing inbox. Those are static repositories you go to when you need to look something up. A hub is people whose job is to actively move learning around the team.
Here’s what a working hub does in practice.
Stays current and atomizes the learning. Hubs translate what’s new and what’s working into bite-sized, role-specific content that lands in the flow of work. Think a two-minute Loom in a Slack channel, not a wiki page nobody opens.
Holds office hours and pairs people. Hubs facilitate live, hands-on experience. A hub member pairs a marketer with AI fluency with a marketer with business context. The work that comes out is better than either could build alone.
Maintains a usable knowledge engine. When engineering firm iMBrace built theirs, they cut their information-search time in half. That’s the number worth paying attention to. The repository is alive, queryable in natural language, and refreshed continuously by the hub itself.
Measures where AI is and isn’t earning its keep. Hubs track what’s working and surface that pattern back to leadership. This is the piece most marketing AI center of excellence job descriptions are missing entirely.
When did your team last build something that flowed AI learning between members on purpose, rather than hoping it would happen at the coffee machine?
Fresh job market data suggests marketing is starting to figure this out. According to a recent piece by Carilu Dietrich, senior marketing AI roles are growing fast under names like head of marketing AI, marketing AI center of excellence lead, and senior director of AI projects. Related GTM engineer postings on LinkedIn more than doubled in six months, from roughly 1,400 in mid-2025 to more than 3,000 in early 2026. Marketing is inventing the hub in real time and giving it a different name.
The teams I see getting it right scope the role correctly. They define the hub lead’s job around moving learning, pairing people, and making the team smarter on purpose. That’s a different job description from “enforce AI standards and police prompt quality,” which is where most of these roles are landing right now.
Where this is headed
The marketing teams that solve workslop in the next 12 months will be the ones that build the connective layer that carries learning between people. When one marketer figures something out, the rest of the team should be using it by the end of the week. That’s the real fix. The efficiency follows. It always does.
(Source: MarTech)
